Fractals and Fractal Architecture

See
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/273457875
Fractals and Fractal Architecture
THESIS · JANUARY 2003

DOI: 10.13140/2.1.3850.9606
CITATIONS READS
8 24
1 AUTHOR:
Wolfgang E. Lorenz
TU Wien
33 PUBLICATIONS 18 CITATIONS
SEE PROFILE
Available from: Wolfgang E. Lorenz

Retrieved on: 04 February 2016
DIPLOMARBEIT
“Fractals and Fractal Architecture”
ausgeführt zum Zwecke der Erlangung des

akademischen Grades eines Diplom-Ingenieurs
unter der Leitung von
O.Univ.Prof. DI Dr.techn. Georg FRANCK-OBERASPACH

Institutsnummer E272
“Institut für EDV gestützte Methoden
in Architektur und Raumplanung”
eingereicht an der technischen Universität Wien
Fakultät für Architektur und Raumplanung
von
Wolfgang E. Lorenz
Matr.Nr. 9025694
Tivoligasse 7-9/2/21; 1120 Wien
Wien, am
Unterschrift
0 Contents ‘Fractals and Fractal Architecture’
0 Contents
1 Introduction page 5
1.1 Mandelbrot – Fractals – Theories and Explanations page 5
1.2 An Overview page 6
2 Fractals – A Definition page 8

2.1 What is a Fractal? page 8
2.2 Characteristics page 8
2.2.1 Characteristics – A Fractal is Rugged page 8
2.2.1.a Coastline page 8
2.2.1.b Border page 9
2.2.1.c Richardson page 9
2.2.1.d Architecture page 10
2.2.2 Characteristics – A Fractal is Self-Similar page 10
2.2.2.a Self-Similar and Self-Affine page 10
2.2.2.b Coastlines page 11
2.2.2.c Self-Similar Structures in Architecture page 11
2.2.2.d Self-Similar Structures and Cities page 12
2.2.2.e Other Disciplines page 12
2.2.3 Characteristics – A Fractal is Infinitely Complex page 13
2.2.4 Characteristics – A Fractal is Developed through Iterations page 13
2.2.5 Characteristics – A Fractal Depends on Starting Conditions page 14
2.2.6 Characteristics – A Fractal is Common in Nature page 14
2.3 Influences page 15
3 Different Fractals page 16

3.1 The ‘true’ Mathematical Fractals page 16
3.1.1 Cantor Set page 16
3.1.2 Sierpinski Gasket page 17
3.1.3 Koch Curve page 17
3.1.4 Minkowski Curve page 18
3.1.5 Peano Curve page 18
3.2 Chaotic Fractals page 19
3.2.1 The Mandelbrot Set page 19
3.2.2 The Julia Sets page 20
3.3 IFS – Iteration Function Systems page 20
3.4 DLA Models – Diffusion-Limited Aggregation Model page 22
3.5 L-Systems – Lindenmayer-Systems page 22
3.6 Midpoint Displacement Method page 23
3.7 Strange Attractors page 23
4 Dimensions page 24
4.1 Euclid page 24
4.1.1 Euclidean and Fractal Geometry page 24
4.1.2 Different Dimensions page 24
4.2 Fractal Dimension page 25
4.2.1 Self-Similarity Dimension ‘Ds’ page 25
4.2.2 Measured Dimension ‘d’ – Structured Walk Method page 26
4.2.3 Box-Counting Dimension ‘Db’ page 27
4.3 Conclusion page 28
Vienna, 2002 – page 2– Wolfgang E. Lorenz

contents ‘Fractals and Fractal Architecture’
5 Fractals and Architecture page 29

5.1 Visual Perception and Fractal Range page 29
5.1.1 Changing of Dimensions page 29
5.1.2 The Scale-Range of a Wall page 30
5.1.3 The Scale-Range of an Elevation page 30
5.1.4 Conclusion page 30
5.2 Topology page 31
5.2.1 Information and Memory page 31
5.3 The Box-Counting Dimension and Architecture page 32
5.4 Symmetry page 33
5.4.1 Symmetry and Mutation page 33
5.4.1.a Cardiff Opera page 34
5.4.2 Symmetry of Fractals page 34
5.4.3 Symmetry of Architecture page 34
5.5 Fractal Characteristics in the History of Architecture page 35
5.5.1 The Time page 35
5.5.1.a The Structure page 36
5.5.2 Gothic page 36
5.5.2.a One Style page 36
5.5.2.b Fractal Characteristics of the Gothic page 37
5.5.2.c Ribs and Tree-Branches page 37
5.5.3 Baroque and Rococo page 38
5.5.4 Predecessor Buildings page 38
5.5.4.a Chartres page 38
5.5.4.b Templo Expiatorio de la Sagrada Familia in Barcelona page 39
5.5.5 Modern page 40
5.5.5.a Le Corbusier – Purism page 40
5.5.5.b Nature and Construction page 41
5.5.5.c Euclidean and Fractal Concepts page 41
5.5.5.d De Stijl page 42
5.5.5.e Conclusion page 42
5.5.6 Organic Architecture page 42
5.5.6.a Frank Lloyd Wright page 42
5.5.6.b Structural Similarity page 43
5.5.6.c The ‘Unity Temple’ page 43
5.5.6.d Eero Saarinen page 44
5.5.7 Contemporary page 44
5.5.7.a Zvi Hecker page 44
5.5.7.b Greg Lynn page 45
5.5.7.c Bernard Tschumi page 45
5.6 The Influence of the Surrounding page 45
5.6.1 Material page 46
5.6.2 Indigenous Buildings page 46
5.6.2.a Translation page 47
5.6.3 Landscape and Building page 47
5.6.3.a Kind of Surrounding page 47
5.6.3.b Two Different Regions page 48
5.6.3.c Including the whole Environment page 48
5.6.3.d Summary page 49
5.7 Analysis page 49
5.8 Helping Tool in the Stage of Planning page 49
5.8.1 Midpoint Displacement page 50
5.8.1.a Fractal Rhythms page 50
5.8.1.b Visually Distinguishable Size page 51
5.8.1.c Some Examples of Fractal Rhythms page 51
5.8.2 IFS in Architecture page 52
5.8.3 Mathematical Fractals page 52

contents ‘Fractals and Fractal Architecture’
6 Fine Arts and City Planning page 53

6.1 Fine Arts page 53
6.1.1 Description of Two Paintings page 53
6.1.1.a Piet Mondrian page 53
6.1.1.b Jackson Pollock page 54
6.1.1.c Conclusion page 54
6.2 Cities page 55
6.2.1 Cities and Fractal Geometry page 55
6.2.2 Brief History page 55
6.2.2.a Modern Grid-Planned Cities page 57
6.2.2.b Natural Growth – Eero Saarinen page 57
6.2.3 Naturally Grown Cities page 58
6.2.3.a Form, Structure and Hierarchies page 58
6.2.3.b Growth Models page 59
6.2.3.c Urban Boundaries page 60
6.2.4 Helping Tool page 61
6.2.4.a Mathematical Fractals page 61
6.2.4.b Curdling page 61
7 Problems with Measuring page 63

7.1 The Size and the Quality of an Image page 63
7.2 The Range of Box-Size page 64
7.3 Starting Points page 64
7.4 The Slope of the log-log Graph page 65
7.5 Programs page 65
8 Pictures page 66
9 Statistics page 121

9.1 The Data-Sheet page 121
9.1.1 The Aim page 122
9.1.2 The Dimension page 122
9.1.3 The Data page 122
9.2 The Evaluation page 123
9.2.1.a I) House Types – Dimension – Floor Plan – ‘Benoit’ page 123
9.2.1.b II) House Types – Dimension – Floor Plan – ‘Fractal Dim. Calc.’ page 124
9.2.1.c III/IV) House Types – Dimension – Elevation page 125
9.2.2.a I/II) Furniture – Dimension – Floor Plans page 125
9.2.3.a I/II) Tiles – Dimension – Floor Plans page 125
9.2.4.a I) Quality – Dimension – Floor Plans – ‘Benoit’ page 126
9.2.4.b II) Quality – Dimension – Floor Plans – ‘Fractal Dim. Calc.’ page 126
9.2.5.a II) Rectangularity – Dim. – Floor Plans – ‘Fractal Dim. Calc.’ page 126
9.2.6.a I) Roofline – Dimension – Floor Plans – ‘Benoit’ page 127
9.2.6.b II) Roofline – Dimension – Floor Plans – ‘Fractal Dim. Calc.’ page 127
9.2.7.a III/IV) Smoothness – Dim. – Elevation page 127
9.2.8.a Variables with no Significant Influence on the Dimension page 128
9.3 The Farmhouses page 128

1 Introduction ‘Fractals and Fractal Architecture’
1 Introduction
“... Clouds are not spheres, mountains are not cones, coastlines are not circles,
and bark is not smooth ...”[01].
This quotation by Mandelbrot shows that the Euclidean geometry – the perfect ‘clinical’
shapes of cones, pyramids, cubes and spheres – is not the best way to describe natural
objects. Clouds, mountains, coastlines and bark are all in contrast to Euclidean figures not
smooth but rugged and they offer the same irregularity in smaller scales, which are some
introduction
important characteristics of fractals – see chapter ‘2.2 Characteristics’.
As the following pages indicate, fractal geometry, in opposition to Euclidean geometry,

offers better methods for description or for producing similar natural-like objects respectively.
The language in which it is expressed is called ‘algorithms’, by which complex objects like a
fern or a cloud can be reduced to simpler formulas or transformation rules respectively.
Fractals can be found everywhere from coastlines, border-lines and other natural rough
lines to clouds, mountains, trees, plants, ... and maybe also in architecture. The following
chapters explain what a fractal is in general and how fractals can be used for architectural
analysis and in the stage of planning. Fractals are used as a helping tool for explanation in
many fields ranging from medicine to economy. From this point of view fractals should not be
excluded from architecture.
1.1 Mandelbrot – Fractals – Theories and Explanations

"... Fractals will make you see everything differently. ... You risk the loss of your
childhood vision of clouds, forests, galaxies, leaves, flowers, rocks, mountains,
torrents of water, carpets, bricks, and much else besides".[02]
The computer-scientist Benoit Mandelbrot introduced the word ‘fractal’[03] in the year 1975
to describe irregular, not smooth, curves. Fractal geometry in general has become more and
more popular since Benoit Mandelbrot’s book ‘The Fractal Geometry of Nature’ was published
for the first time in 1977 and the ‘Mandelbrot set’ was defined in the year 1980. Since that time
many of those computer-pictures and ‘monster curves’, as Benoit Mandelbrot called curves
Mandelbrot
with unusual characteristics such as the Koch curve, have been published – some of them will
be introduced in chapter ‘3 3 Different Fractals’. They were mostly created with a mathematical
background, but they can also be found in art because of their beautiful interesting
appearance.[04] In the eighties of the 20th century the ‘Mandelbrot set’ could be found in many
scientific journals last but not least because of its beauty – this kind of visualizing was,
however, only made possible by the quick development of the computer. Subsequently today
it is no problem for any of us to produce classical and natural looking fractals quickly and easily
on the computer at home.
new geometry
Of course it is dangerous to think that we know enough about fractal geometry and the
mathematical background so that we can start to use this ‘new’ geometry in every field of our
life. But, are we not now – with all the background of our society – in a position where we can
deal with all the questions arising from the use of these new scientific theories and is this not

main part ‘Fractals and Fractal Architecture’
more important than putting all the priority on finding the final result at the moment. To use new
theories first means to define how they can help in a specific field and then to find applications
in the next step. These days we can already look into other possible fields of application of
fractal geometry: natural science, medicine, market analysis, manufacturing, ecology – and
architecture?
In any case architects and city-planners already use concepts and phrases for analysis
vocabulary
which come from other fields, such as the biological metaphor: the city as an organism with the
‘heart’ for the central business district, the ‘lung’ for the green space, the ‘arteries and veins’ for
the hierarchy of communications routes[05] – in this connection fractal geometry might provide
a new vocabulary which explains the complexity of a city in a better way.
1.2 An Overview
We can describe mountains, clouds, trees and flowers by models consisting of
simpler geometric forms based on Euclidean geometry, for example using net
models in CAD, but are they exactly what nature is?[06]
Euclidean – fractal geometry

The first two chapters below give an introduction to fractals and fractal geometry in a more
general way, listing characteristics and explaining some examples. Then one chapter follows
about the differences between Euclidean and fractal geometry and their expressions in the
Euclidean and fractal dimension, introducing and explaining some measuring-methods of the
fractal dimension.
Until a short time ago scientists described nature through so called ‘smooth’ continuous
mathematics, which is the mathematics of smooth forms such as lines, curves and planes and
which is expressed in the language of Euclidean geometry. The ‘new’ science of complexity[07]
does not try to simulate any more the rugged character of nature through smooth forms but it
deals with the raggedness of the structure itself – this field of mathematics is expressed in the
language of fractal geometry: ‘The whole is more than its parts’.
“The fractal new geometric art shows surprising kinship to Grand Master
paintings or Beaux Arts architecture. An obvious reason is that classical visual arts,
like fractals, involve very many scales of length and favor self-similarity".[08]
Chapter ‘55 Fractals and Architecture’ to chapter ‘6 Fine Arts and City Planning’ deal with the
question if fractal geometry can be used in architecture and for the analysis of cities or at least
with the question in which ways further research can be done – it lists the possible applications
of fractal-theory in fine-art, city-planning and architecture, with regard to analytical, formal and
practical aspects.
Fractal geometry may help us to understand and analyze complexity that can be found in
towns of the Middle Ages but also in cathedrals and other man-made objects up to these days.
complexitiy
It may also help us to transfer this complexity, which also arises from the development over
time, to newly planned cities and buildings – cities and buildings may then also be reduced to
simpler algorithms: the automatic architect?[09]

Among other themes one part of chapter ‘5 5 Fractals and Architecture’ deals with the so
organic architecture
called ‘organic’ architecture. In this context, natural structures, explained through fractal
geometry, could serve as easy-to-use models for a new ‘organic’, natural-like architecture.
The underlying structure, the character of a building will then follow the principles of fractal
geometry that can be found in natural elements. This results in a connection between
architecture and nature not only in a formal respect – e.g. with Antoni Gaudí[10] – but also
in the complexity of forms and in its character of self-similarity, irregularity and roughness
– e.g. with Bruce Goff.
Another section of this chapter deals with the question of the quality of buildings, which
quality of buildings
means asking why one building is regarded as ‘good’ architecture while the other is regarded
as ‘bad’ architecture. It is clear that everyone has his or her personal access to architecture
and evaluates buildings in a different manner. But there are buildings, which are more
interesting in general than others, or there are buildings, which are more noticeable to most
people than others – e.g. the acceptance of Gothic, which is related to fractal geometry, in
contrast to the resentment towards modern buildings, which are related to Euclidean
geometry. Subsequently the idea arises that maybe it is complexity that modern buildings lack.
Pictures with additional explanations to all previous chapters are concentrated in one
section, chapter ‘8
8 Pictures’, for giving a better compact view.
Chapter ‘77 Problems with Measuring’ deals with measuring problems arising from the use
of different computer-programs for calculating the box-counting dimension introduced in
chapter ‘44 Dimensions’. Finally, this box-counting dimension is used in chapter ‘9
9 Statistics’ to
get a classification of elevations and ground plans of buildings and cities. In this context the
fractal dimension is applied as an indication for roughness.

2 Fractals – A Definition ‘Fractals and Fractal Architecture’
2 Fractals – A Definition
The best way to define a fractal is through its attributes: a fractal is ‘rugged’, which means
that it is nowhere smooth, it is ‘self-similar’, which means that parts look like the whole, it is
‘developed through iterations’, which means that a transformation is repeatedly applied and it
is ‘dependent on the starting conditions’. Another characteristic is that a fractal is ‘complex’,
but nevertheless it can be described by simple algorithms – that also means that beneath most
natural rugged objects there is some order.
2.1 What is a Fractal?

“Fractals are objects of any kind whose spatial form is nowhere
why characteristics
smooth, hence termed ‘irregular’, and whose irregularity repeats itself
geometrically across many scales”.[01]
In fact there are so many different types of fractals, some of which will be introduced in
chapter ‘3
3 Different Fractals’, that it is not possible to give one definition for all of them.
Besides, when we are talking about fractals in general we should never forget that there are
many which have not been found yet. Considering this circumstance, it is more useful to
describe some of their characteristics.
2.2 Characteristics
characteristics – a summary
"... the world is chaotic, discontinuous, irregular in its superficial physical form
but ... beneath this first impression lies an order which is regular, unyielding and of
infinite complexity."[02]
2.2.1 Characteristics – A Fractal is Rugged
2.2.2 Characteristics – A Fractal is Self-Similar
2.2.3 Characteristics – A Fractal is Infinitely Complex
2.2.4 Characteristics – A Fractal is Developed through Iterations
2.2.5 Characteristics – A Fractal Depends on Starting Conditions
2.2.6 Characteristics – A Fractal is Common in Nature
2.2.1 Characteristics – A Fractal is Rugged

length-measurement
2.2.1.a Coastline
Benoit Mandelbrot, the ‘father’ of the popularity of fractals today, introduced fractal
geometry by the question of how long the coastline of Britain is. This question of length seems
to be very trivial but nevertheless there is more than one possible answer.
To get a better understanding of the problem of length-measurement first of all let’s have a
closer look on how the coastline of Britain may be measured: for measuring the length of the
outline it is possible to use a certain map with a certain scale, following the coast by a ruler.

When we repeat this way of measuring on another map of a smaller scale, the outline turns out
to be longer. The reason lies in the fact that the latter map will offer more details, that is sub-
bays, inlets, cliffs and rocks, whose circumferences are included in the measurement which
finally leads to the increased length, see picture 01. So the coastline may be longer or shorter,
depending on the scale of the map we are using for measurement.
Looking down on the coastline from a great distance, out of an airplane, we will
recognize the character of the border on principle – if its rough or smooth –, but
we will not see all the small inlets which will come up when we are closer to the
coastline, for example when walking along the beach.
Nevertheless there are some limits to this length-measurement. One limit is the size of an
atom as a physical border – theoretically, without any limit, the length of the coastline would
reach infinity at an infinite small scale because of the infinite number of inlets. The other limit
is the correct definition of the coastline, that is where the exact border between water and land
limits
is and at which time it should be measured, at high tide or low tide. Consequently, it may then
be better to choose a rock of about 20 meters length as the lower limit instead of an atom. This
value also arises from the fact that regions of coastlines have been cultivated by man –this
has turned them into smoother parts. Therefore, to avoid falsifications, smaller scales have to
be excluded from measurement.
2.2.1.b Border
Another example of scale-dependence is the length-measurements of borders between
length-measurement
countries. The problem of measuring them has been known for a long time and in most cases
the deviations do no harm: e.g. the circumstance that the length of the border between Spain
and Portugal is given differently by the officials of these two countries has no consequences.
The difference results from the fact that the official maps of Spain have a bigger scale than
those of Portugal.[04] Here we find the same phenomenon as with the coastline above – the
maps of Portugal show more edges and corners, which mean that the border is longer than
on the maps of Spain. The length given in the Spanish encyclopedia is 987 km and that in the
Portuguese one is 1.214 km.[05]
Boundaries of cities are similar to borderlines of countries and outlines of coasts – e.g. the
political border, the changeover from the city to the natural environment, from high-density built-
up areas to low-density areas, from regions of dwelling houses to regions of one-family houses.
cities
Nevertheless, there are again limits as with the coastlines above – like the definition of the
border, that is defining its way and the limit by the smallest and largest scale. In addition to that
the fact that various measuring methods lead to different results is valid for all measurements
of fractal dimensions, which will be described in chapter ‘4.2 Fractal Dimension’.
2.2.1.c Richardson
As early as in the year 1961, Lewis Fry Richardson examined the growth rate of the length
for different curves such as coastlines and borderlines, by replacing the original curve by a
polygon consisting of equal-sized lines ‘e’ – the unit length. For each curve and for a certain ‘e’
he got an overall length L(e) through the approximation formula L(ee)~l l*ee(1-D). – That means

that Richardson put the total length L(e) in relation to the unit length ‘e’ by using two constants
l and ‘D’: ‘D’ depending on the coastline and its roughness.
For comparison, he then decreased the unit length ‘e’ for each curve and measured once
more, which led to another greater length L(e), and so on. The results of measurement were
put into a graph of log(unit length) across log(overall length)[06]. The curves of each measured
borderline of the graph led to a certain gradient angle, which was interpreted by Mandelbrot
as amounting to about ‘1-D’: Mandelbrot called ‘D’ the fractal dimension of the curve, e.g. for
a circle the gradient angle turns to zero which equals D=1.
comparison of lengths
unit length 300 unit length 25
unit length 100
As unit-lengths lead to different total-lengths, borders, coastlines and other

fractal curves cannot be compared by their length. Firstly, because one ‘unit-
length’ for all measurements would have to be defined, which has not been done
yet, and secondly, nature also exists without man, which means that a typical man-
defined unit-length would place man above nature.[07] Therefore the fractal
dimension ‘D’ will be a better parameter for comparison.
2.2.1.d Architecture
Facades often display some kind of roughness – for example let’s think of
window-frames, the distribution of windows and doors, the character of bricks and
other materials, the structure of the roof and the wall.
2.2.2 Characteristics – A Fractal is Self-Similar
2.2.2.a Self-Similar and Self-Affine
Fractals are always self-similar, at least in some general sense – what does that mean?
That means that on analysis of a certain structure will bring up the same basic elements on
different scales. For example, details of a certain coastline look like larger parts of the whole
curve; the characteristic – the irregularity – of this natural form remains the same from scale
to scale. In this way fractals can also be described in terms of a hierarchy of self-similar
components – e.g. trees and branches or town-, district- and local-centers.
Any structure is self-similar if it has under-gone a transformation in which the proportions

of the structure have all been modified by the same scaling factor. The new shape may be
smaller, larger, rotated, and/or translated, but its shape remains similar, which means that the

relative proportions of the shapes’ sides and the internal angles remain the same[08] – these
transformations can be produced by a reduction-copy machine as shown in picture 02. Fractals
self-similarity
produced by self-similar transformations are ‘true’ fractals, the underlying algorithm is the same
from scale to scale – zooming into such a fractal shows an object, which is the same as the
whole. These ‘perfect’ fractals may produce objects which look similar to nature, but there is
mostly something missing – the factor of random. Nevertheless they can be used as a first
approach to nature instead of Euclidean objects – remember: ‘... trees are no spheres’.
If a transformation reduces an object unequally in one or the other way, then the
transformation is referred to as a self-affine transformation. In a self-affine transformation the
internal angles of the shape and/or the relative proportions of the shape's sides might not
remain the same – these curves are not exactly self-similar. For example parts of a ‘natural’
fern or a snowflake are not exactly a copy of the whole as it is true for the computed examples
self-affinity
of fractals, but nevertheless the parts look very much like the whole – this is called a statistical
self-similarity because on average parts look equal.[09]
Some examples of statistically self-similar structures in nature are the large and
small branching structures of a tree, the bays and inlets of a coastline, weather
fluctuations through time. But also man-made structures can be statistically self-
similar such as the average of the Dow Jones, which shows similarities between
hourly, monthly and yearly fluctuations.[10]
2.2.2.b Coastlines
self-similarity of coastlines
Coastlines in reality look similarly rugged no matter from which distance we observe them.
That was also one important circumstance that Benoit Mandelbrot found out when he analyzed
the coastline of Great Britain: A certain part of the curve always looked similar to the whole
curve no matter on which scale he searched.[11] For example Norway has a lot of fjords, which
means that its coastline is very rugged – zooming in on the fjords will offer a similar roughness.
That the coastline of Norway is very rugged means that we will more quickly
find a greater number of details from scale to scale and the length increases more
quickly than in the case of a smooth coastline. The quickness the length
approaches ‘infinity’ is a characteristic for the examined coastline.
On looking at the border of a Euclidean object such as a circle, something different

happens. As we did with the coastline we zoom in infinitely on the curve, but instead of finding
circle
similar sections the zoomed part turns into a straight line. In contrast to that we would never
find a straight line on the coastline but always a rugged part of the curve.
2.2.2.c Self-Similar Structures in Architecture

Castel del Monte
Castel del Monte: This is an early architectural attempt at self-similarity. The building
has been erected over an octagon, with octagonal towers on its edges. The same form can
be produced by putting the octagon in the copy machine of picture 02 with the input:
‘Reduce the octagon by a certain factor 1 to X and put it on the edges of the previous
octagon’. This could be repeated more often. In fact, what we see here is the first iteration
of this expression – see picture 03.

Bruce Goff: Some of Bruce Goffs buildings contain characteristics of fractal geometry, for
example the Eugene Bavinger house, near Norman, Oklahoma, built in 1950. The floor plan
describes a curve that shows a form like the self-similar Cephalopode Nautilus – the units of
the Nautilus follow the structure of a logarithmic spiral curve. This curve is called self-similar
Bruce Goff
because the angles of the tangents are equal in all points – see picture 04.
Another example of fractal architecture by Bruce Goff is the Joe Price Studio in Bartlesville,
Oklahoma from 1956. All shapes include equivalent triangles on different scales; the angles
are similar from scale to scale – 60 degrees or a multiple of it. The ‘meeting-area’ is e.g.
hexagonal, the walls of the music-room are three-cornered and the ceiling of this room consists
of a three-cornered decoration for better sound quality through reflection – see picture 04.[12]
Facade: Looking at a building from some distance, we will get a first impression of it.
Coming closer, it depends on the respective building itself if some more details of interest
appear. As an example the view of Robie house by Frank Lloyd Wright will offer more and
F. L. Wright
more details on a lower scale with a repetition or variation of forms respectively. The building
‘remains’ interesting – and in addition to that the repetition helps us with regard to orientation
and classification. In contrast to Frank Lloyd Wright, the smooth modern buildings of the
twenties and sixties often lack detail on lower scales – as we will see later on they also have
a lower fractal dimension.
Architecture: For more examples of self-similarity from scale to scale and other fractal
characteristics in architecture – from ground plans to facades – see picture 05.
2.2.2.d Self-Similar Structures and Cities
Cities: Self-similar patterns can also be found in cities in a more abstract respect. If we
compare the function of rooms of a dwelling or a one-family house – living, working, walking,
cities
relaxing –, that is to say the way of dividing up space, with a quarter and the whole city we will
find analogies in areas for working, living etc. Functions are repeated on different scales. But
also the repetition of ways observed over years between the separate functions inside a
dwelling within a day or ways in the city, e.g. to and from work, can be seen as fractal concepts.
Africa: The site plan of a village in Cameroon/Africa is based on a circle, which is then
village
repeated on the smaller scale of the ground plans of the cottages – see picture 06.
2.2.2.e Other Disciplines

Self-similarity is also found in other disciplines – the structure of our memory for example
is also fractal. A certain word conjures up a complex network of images in connection with
human memory
that word. It is not the object or similar objects that come up, but events and associations, for
example a certain mountain tour for the word ‘mountain’. By that a set of memories comes up
to one’s mind, people that were present at the tour and other thoughts in connection with them.
The word does not bring an abstract image to one’s mind but a complex set of interrelations
to events and persons. Hardly any boundary can be drawn around these interrelations, which
is also true for the border of the Mandelbrot set. The network for a certain word is of course
different from person to person, because of individual experiences, but this is not so important.
The different networks have enough similarities for communication. For conversation and
books the fact that different networks exist is getting more and more of a problem because the

information increases and so the nets that are conjured up and as a consequence also the
interpretation of the meaning differ more and more.
2.2.3 Characteristics – A Fractal is Infinitely Complex
Fractals are highly complex, that means zooming in will bring up more and more
details of the object, a characteristic that continues until infinity.
Julia set
Chaotic fractals: Already in the 20ies of the 20th century the French mathematicians
Gaston Julia and Pierre Fatou concerned themselves with the question of fractal geometry.
Both examined what would happen to a point ‘Z’ of the complex number-plane if the
transformation f(z)=Zn2+C was repeatedly applied to it.[13] Gaston Julia discovered that
calculating the function repeatedly might deliver unforeseen ‘chaotically’ results.
It was only in the 70ies that Mandelbrot[14] could show the results of the formula as a
picture, which needed the high capacity of computers. The Mandelbrot set is similar from scale
to scale, which means zooming closer to the details there will always come up new parts
Mandelbrot set
looking similar to each other and sometimes to the whole – see picture 07. The only limits are
limits of capacity and, resulting from this, rounding mistakes by the computer, but also limits of
the visual medium.
For further information about Julia sets and the Mandelbrot set see chapter
‘3.2 Chaotic Fractals’.
2.2.4 Characteristics – A Fractal is Developed through Iterations
Self-similarity, as described before in this chapter, can be produced by iterations, which
iteration
means that certain kinds of formulas or geometric principles are repeated on the previous
result of the calculation or drawing respectively. Examples for geometric rules make up the
fern and the Koch curve; those for fractals based on a mathematical equation produce the
Mandelbrot set.
A film-camera and a television viewer in a dark room can illustrate feedback,

which is a phenomenon that is produced by iterations – iterative processes are a
main source for complexity. The distance should be about one meter and the cut,
which is filmed, should be a little bit larger than the screen. What we then can see
is an endless picture of the screen on the screen on the screen ... Now the camera
is turned a little bit out of the axes. At one point the screen begins to flash. The
feedback
light, which is sent out by the glowing phosphor-layer of the television viewer, is
met by the lens of the camera; this produces flickering electric streams. A cable
leads the electric signal to the screen where it makes some more phosphor glow.
Then the next iteration starts. Depending on the adjusting knobs, like for
brightness, different things may happen. Some patterns, which can then be seen
on the screen, are really constant, even if a hand is held between the screen and
the camera or the light is turned on and off, which means that similar pictures will
again be produced only after a short time.[15]

2.2.5 Characteristics – A Fractal Depends on Starting Conditions
Little differences with regard to the starting conditions may cause great
differences in the results. The reason for that lies in the circumstance that for
fractal structures always the same rules are repeatedly applied.
Chaos and fractals: In connection with certain non-linear systems the word ‘chaos’ does
not mean disorder or confusion – because in this chaos there exists some order and regularity
and it can also be controlled in some way –, but the irregular not foreseeable development in
time of such systems. Chaotically systems can be represented by fractals, which also means
that they have the same characteristics. One characteristic is the dependence on and
chaos
sensitivity of changing the starting values and conditions, another is self-similarity. In chaotic
systems even small changes of the starting values result in great differences in the end. This
phenomenon has been called the ‘butterfly-effect’[16].
In chaotic systems natural laws truly exist but the development of the system
cannot be forecast, therefore the phenomenon is called the ‘determined chaos’.[17]
The behavior of a system can be analyzed by repeating the experiment with the same
starting values, under the same conditions so that the same results may be found. This leads
us to the principle of causality. If the same causes have the same effects this is called a ‘weak’
causality. But mostly it is only possible to get similar starting conditions and not the same ones,
which leads to a ‘strong’ causality – similar cause lead to similar effects. The weather model by
Lorenz is no such system – in this case similar causes have different effects. Such systems had
causality
been regarded as exceptions, but meanwhile they seem to be the normal case.[18]
Up to now mathematicians and physicists have found out that the ‘butterfly
effect’ exists almost everywhere and that a ‘strong’ causality can only be found at
the beginning. The world is balanced between total order, with simple laws, and
total disorder. If architecture can be seen as a mirror of time and society, this new
knowledge has to be translated into buildings and city planning.
2.2.6 Characteristics – A Fractal is Common in Nature
Many objects show fractal structures and can be reproduced through fractal geometry such
as the cluster of galaxies, the roots of trees, the crater landscape of the moon and plants.
nature
It is a truism that most of the following analogies between fractal geometry and
architecture or nature are not based on the pure definition of fractals given above,
but in general they have similar characteristics. For better differentiation this
second kind of fractals will be called ‘natural’ or ‘general’ ones in contrast to the
‘true’ ones.

2.3 Influences
differences natural – ture fractals

There is one important fact about the group of ‘general’ fractals namely the natural
development. This means that for the growth of natural but also for artificial objects many
additional influences have to be thought of. Thus a tree or a fern can be produced by fractal
geometry but these pictures nevertheless offer some differences in respect to their natural
brothers. A tree standing alone on a hill for example is influenced by the wind blowing there,
which forms the tree in one typical form: branches are only to be found on the side turning
away from the direction the wind blows. Other influences may be soil and water conditions,
kinds of plants nearby and animals.
Nevertheless ‘true’ fractals can produce typical natural and man-made forms,
but only under sterile conditions. If some random factor is added then the resulting
objects come nearer to the ‘real’ world.
Likewise the development of cities underlies a couple of influences like natural barriers
such as hills and rivers, but also man-made ones such as roads leading to other towns,
important industrial areas and green-zones of which the growth of the city reacts. The same is
true for elevations and even ground plans of buildings that react to the surrounding no matter
cities
whether it is man-made or natural.
From that follows that if we know the underlying algorithm of any object – under
sterile conditions – and if some mechanisms for simulation of certain influences
are added, we may determine future developments of e.g. the growth of a city.

3 Different Fractals ‘Fractals and Fractal Architecture’
3 Different Fractals
Every natural thing around us is a fractal structure in principle, because smooth
lines and planes only exist in the ideal world of mathematics. Beside that theoretically
any system, which can be visualized or analyzed geometrically, can be a fractal.
This chapter gives an introduction to some different kinds of fractals like the so-called
‘true’ mathematical fractals, to which the Cantor set belongs, and the ‘chaotic’ fractals, with
the Mandelbrot set being an example. Beside that some other methods of creating fractals
such as the iteration function systems, the DLA model, the L-system and the Midpoint
displacement method will be introduced. The form of strange attractors as a connection to
deterministic chaos also offers fractal characteristics and will be described at the end of this
chapter. The one or other type of fractal may help in creativity, analysis, comparison,
construction, organization and other questions arising in architecture.
3.1 The ‘true’ Mathematical Fractals
simple rules
The development of this kind of fractals consists of simple rules – a starting image, the so-
called initiator, is replaced by another image, the so-called generator. But nevertheless they
are very complex and always strictly self-similar: it does not matter which part we analyze, it
always looks exactly like a scaled down copy of the whole set. The tools to create such fractals
are called iteration and feedback: Iteration means that the procedure is repeated based on the
result of the previous step.
Fractals in the mathematical sense only exist at the limit point of an infinite number of
generation steps. Thus a part will be an exactly scaled down copy of the whole if both, the
part and the whole, contain infinite small sections. Therefore they can only be found in
theory: infinite steps are only reached after infinite time. This also means that the pictures
limited range
of the fractals shown on the following pages are only approximations but they will already
offer the characteristics of fractals after two or three steps. That also leads us to a more
abstract definition:
Fractals are objects, which are self-similar at least in a more general sense up
to a certain scale – this is also true for coastlines and elevations of buildings, as I
have mentioned before.
3.1.1 Cantor Set[01]
For producing the Cantor Set the initiator, a straight line of a certain length, is replaced by a
construction rule
generator consisting of two lines, each of the length of 1/3 of the initiator, in such a way that the
new lines are located in each case at the end of the initiator. From that an open middle interval
of the same length as the lines of the generator emerges but this ‘hole’ does not include its end
points – these points belong to the two outer parts, marked 1 and 2 in picture 08. This geometric
rule is repeated again with the two new lines, which leads to four lines and so on.
The process of constructing the Cantor Set is called coagulating, with the mass
of the middle third flowing into the right and left section.

After infinite steps the result is an infinite number of clustered points with an infinitely high
density, which are situated along a line. The Cantor Set is one example of curves that Benoit
Mandelbrot called ‘monsters’ because of their unearthly characteristics. The Cantor set[02]
fractal characteristics
demonstrates very clearly two important features that fractal structures have in common:
1. Self-similarity from the large to the small scale: Any part of the Cantor Set is
an exactly scaled down copy of the entire Cantor Set.
2. The clustering of the points in the Cantor Set is similar to natural systems
with a random factor being added – e.g. the distribution of stars displays a
clustering in their layout rather than an even or strictly random distribution of
features.
The distribution of rough diamonds in the earth's crust is quite similar to the distribution of
stars and galaxies in the sky – Benoit Mandelbrot presented such a comparison in his book ‘Die
fraktale Geometrie der Natur’[03]. If we mark every diamond-mine and every place of discovery
galaxies
– old and new one – on a world-map, the density of these marks will look irregular from a great
distance: some are isolated and most are concentrated in some few regions. In these regions
the distribution of diamonds is not regular either. On closer observation, there are again only
some isolated concentrations of diamonds within ‘poorer’ regions, which looks similar to the
whole like a scaled down copy. What we see is a two dimensional translation of the Cantor set.
This sounds quite similar to the distribution of centers or the distribution of

high density of population: both offer a clustered pattern. For example a city
cities
consists of areas of high-density within regions of lower density, and on a world-
map the distribution of cities again looks irregular, with large regions of no
density in between.
3.1.2 Sierpinski Gasket[04]
construction rules
For producing the Sierpinski Gasket, the initiator, an equilateral triangle, is replaced by a
generator consisting of three equilateral triangles, each of the size of half the initiator, in such
a way that the new triangles are located in each case at the three corners of the initiator. In
other words an equilateral triangle is cut out in the middle. This cut-out triangle is half the size
of the initiator and rotated by 180 degrees – the side-points of the triangle are defined by the
midpoints of the sides of the original triangle. The same procedure is repeated for each of the
three new triangles, and so on. The remaining triangles or the set of points that are left after
infinite iterations is called the Sierpinski Gasket. For further details see picture 09.
3.1.3 Koch Curve[05]

construction rules
The initiator of this fractal is again a line, the generator four lines of 1/3 of the initiator.
For their creation, the initiator-line is divided into three equal parts, with the middle part
being replaced by an equilateral triangle of the side length of 1/3 of the initiator – the lower
part of the triangle, however, is taken away. This procedure is then repeated for the four
new lines. After infinite steps the construction leads to the Koch curve – the geometric rule
for this fractal is shown in picture 10 together with a description why this curve belongs to
fractal geometry.

The Koch curve demonstrates how one can get a curve of infinite length. The original line,
the initiator, may have a length equal to one. This line is replaced in the first step by four lines,
which is called the number of pieces N, of 1/3 of the original length, which is called the reduction
factor s. So the resulting length of the ‘new’ curve is L1=N1*s1=4*(1/3)=1.33, which is longer
than the original line L0=1. In the second step four other lines again replace each of the four
lines. N2 is therefore 4*4=16, of a length of 1/3 of 1/3 of the original line so this leads to the
infinite length
second length L2=N2*s2=4*4*(1/3)*(1/3)=16/9=1.78, and so on:
L3=N3*s3=4*4*4*(1/3)*(1/3)*(1/3)=64/27=2.37 ...
Hence it follows that the length of the curve increases by L=(4/3)n, with ‘n’ being
the number of steps, that is iterations and the generator length being 1:
L10=N10*s10=(N1*s1)10=(4/3)10=17.75
L50=(N1*s1)50=(4/3)50=1765780.96
The Koch Island, which consists of three Koch curves placed along the sides of
an equilateral triangle, is a first approach to a snowflake but also to an idealized
city plan, as we will see in chapter ‘66 Fine Arts and City Planning’.
3.1.4 Minkowski Curve
For constructing the Minkowski curve the initiator, a line of e.g. a unity length equal to one,
construction rules
is replaced by a generator consisting of eight lines. These eight lines, each 1/4 of the original
line, are arranged in the following manner: horizontal lines, which are kept in position, build the
first fourth and the last fourth of the original line. The second fourth consists of a line turned
up 90 degrees, followed by a horizontal line and finally by a line moving down 90 degrees
again. The third fourth is constructed by a sequence of lines that is first turned down 90
degrees then moving horizontally and finally turning up 90 degrees again to connect the last
fourth. This rule of construction is then repeated for all eight new lines of the first iteration, 64
lines of the second iteration, 512 lines of the third iteration and so on – see picture 11.
In comparison to the Koch curve the length of this curve grows even faster from one stage
of construction to the next. The length is measured by the equation L=(8/4)n, the generator
length
again being a line with the length of one. After the first iteration the length is given by
L1=N1*s1=8*(1/4)=2, after the second step L2=N2*s2=(8/4)2=4 and after the 10th iteration
L10=N10*s10=(8/4)10=1024.
The resulting Minkowski curve looks like a coastline that turns in and out with
coast
bays and inlets. By random the resulting curve turns out to be even more
realistic – see picture 12.
3.1.5 Peano Curve[06]

rules
The initiator of the Peano curve is once more defined through a straight line and the
correspondent generator consists of nine lines 1/3 of the initiator. The first line of the
generator runs horizontally, the second turns up by 90 degrees, then a horizontal part follows

before the curve turns down again by 90 degrees. The fifth line moves back to the end of the
first line without touching it. The next part of the curve heads down by 90 degrees, followed by
a horizontal part before it goes up again. Finally a line located in horizontal position again
forms the last section – see picture 13.
The Peano curve offers a paradox. It is a curve that fills the surface it is lying on after infinite
iterations. So this curve is at the same time a one-dimensional entity, a line, and somehow also
one- and two dim.

a two-dimensional unit, a plane. The phenomenon we find in this structure is the fact that a one-
dimensional curve, in Euclidean terminology, has a fractal dimension of two.
Some examples of space-filling curves are the distribution of arteries and blood
vessels in the body, the brain, the roots of trees, the pulmonary arteries and others
– they all offer a structure of tree branches that nearly fills its space.
If we interpret the curve as a street put on top of the surface of a city, every part or house
of this city – the two dimensional plane – would be reached via only one way. Using the Peano
street-planning
curve as a street would also mean that the whole plane is taken in by the curve and that getting
at the end of this street would lead along the whole, ‘infinite’, curve which would not be
desirable. But with the aid of fractal dimension, the measure of how much the curve exceeds
its base dimension, that is the measurement of the roughness, one may be able to measure
how many areas of the plane are reached by the curve. This can also be useful for the
planning of streets, where one wants to reach a certain number of houses via one street.
3.2 Chaotic Fractals

Another category of fractals represents the so-called chaotic, non-linear fractals. This fractal
general explanaitions
type is connected with the theory of chaos, and its elements are obtained by a simple
mathematical equation.[07] For visualizing them, each point on the paper or screen is related to a
certain number – e.g. in the case of the ‘Mandelbrot set’ this is a complex number. This number
is then iterated, that means it is used in a formula and the new number resulting from that is then
again used in the same formula, which leads to the next iteration. This sequence of operations is
‘similar’ to the work of the ‘copy-machine’ of linear fractals – with regard to insertion.
The insertion is repeated until the numerical values approach infinity, converge
or fluctuate between several numbers. Depending on the result, the original point
may be colored differently.
3.2.1 The Mandelbrot Set
The Mandelbrot-fractal itself is the picture of the Mandelbrot set – the Mandelbrot set being
the numerical set of the complex numbers for that is valid if, being repeatedly put into the
the plane
formula Z(n+1)=Zn*Zn+C, the absolute value remains a finite number.[08] The plane in which it
is drawn is called the complex plane where each point represents a complex number of the form
C=a+bi – with Z0=0+0i and Zn=X+Yi. The system of co-ordinates of the plane or screen is
defined with the x-axes representing the value of ‘a’ respectively ‘X’ and the y-axes the value of
‘bi’ respectively ‘Yi’ with ‘i’ being the root of -1. Numbers multiplied by ‘i’ are called imaginary
numbers ‘b’, those which are not are called real numbers ‘a’.

For visualizing the set, each pixel of the screen, representing a certain complex number, is
iterated in the formula Z(n+1)=Zn2+C. For each pixel, point, ‘Z0’, mostly fixed in the zero point
Z0=0+0i, remains the same but the value of ‘C=a+bi’ is chosen differently. After a certain
the ruel and the set

number of iterations there are two possibilities, first the iteration for a specific ‘C’ diverges to
infinity or second it does not diverge but approximates a certain number or fluctuates between
several numbers, so-called fix-points. If the iteration remains limited, the analyzed pixel
Z1=Z02+C for a specific ‘C’, is an element of the Mandelbrot set and colored black. If it diverges
to infinity mostly the steps are counted and, depending on how fast it diverges, the pixel is given
a special color. Then the procedure is repeated for another value of ‘C’ and so on.[09]. What is
fractal about that are the infinitely small boundaries between ‘Zn’ being a finite value or jumping
between certain numbers and ‘Zn’ diverging towards infinity – see picture 14.
3.2.2 The Julia Sets
construction rules
The Julia sets use the same formula as for the Mandelbrot set, whereas this time for one
certain Julia set, the value of ‘C’ keeps the same and ‘Z0’ is changing. The behavior of each
starting point Z0 is examined by iterating it in the formula Z(n+1)=Zn2+C, whereby the same
complex plane as for the Mandelbrot set is applied. Finally the examined point belongs to the
Julia set if ‘Z(n+1)’ keeps a finite value. The procedure is repeated for each pixel, containing
different starting values ‘Z0’, of the screen.[10]
For each value of ‘C’ in the Mandelbrot set there exists one specific Julia set, which shows
us that there are unlimited Julia sets as there are infinite points of the Mandelbrot set.[11] Julia
sets for points outside the Mandelbrot set look like dust or a disconnected cloud of points,
Mandelbrot – Julia sets

they consist of infinite loose points. Whereas Julia sets for points being elements of the
Mandelbrot set itself are connected, which means that each point of the real Julia set, mostly
colored black, has a neighboring-point. Those for values of ‘C’ in the heart-shaped main-body
of the Mandelbrot set look like rugged, deformed circles – see picture 14.
How can it be proved easily that a certain set is not interrupted? Gaston Julia found out that
one only has to analyze the behavior of the critical zero-point, Z0=0+0i. The resulting picture
of ‘C’ is continuous if the set of iterations of these points 0, C, C2, (C2+C)2+C does not grow
too extremely.
3.3 IFS – Iteration Function Systems
Iterated function systems, the so-called IFS, again belong to the types of linear fractals like
the true mathematical fractals. They are produced by polygons that are put in one another and
show a high degree of similarity to nature – such as the fern presented in picture 15. The IFS
form the connection between the true mathematical fractals and nature.[12]
The IFS can be illustrated with the Sierpinski Gasket by its insertion-rules, remember the
copy machine of picture 02: a ‘starting’-square is replaced by three squares of half-size, one
insertion
situated on the middle top, the other two on the left and right bottom. In the next step each of
the three squares is replaced by another three shorter squares and so on. It is not the starting
image, in this case a square that influences the resulting picture but the rules itself – see
picture 16. The only difference lies in the detail, zooming in on the object will show the

different starting objects, except after infinite iterations when those single objects turn to
infinitely small ones. In the IFS, the transformation rules can also include rotation, reduction,
enlargement, shear and similar rules, which are described in mathematics as affine linear
transformations – see chapter ‘2.2.2 Characteristics – A Fractal is Self-Similar’. For each
single lens or single insertion rule these transformation rules can be chosen differently.
It will require many iterations, runs of insertions, until the single objects, e.g. the rectangles
of Barnsley's fern shown in picture 15, cannot be identified any more. In the case of the fern
this will take about 50 iterations, which means that 450 rectangles have to be drawn –
N=nm=450, with ‘n’ number of transformations, ‘m’ number of iterations and ‘N’ number of
affine transformatins
elements of the resulting object after ‘m’ iterations.[13] Generating these 50 iterations takes a
lot of time even with the assistance of the computer, therefore another method is chosen to
get the resulting fern, the so-called ‘chaos game’ – see picture 17, an explanation is also
given at picture 09 ‘Sierpinski Gasket’. In that case the four insertion-rules are expressed by
four mathematical affine transformations[14] of the form f(X,Y)=(X(n+1)=A*X(n)+B*Y(n)+E,
Y(n+1)=C*X(n)+D*Y(n)+F). For IFS these transformations must be contractions, which are
reached in the way, that A, B, C, and D remain within the range of -1 and +1. An IFS always
consists of more than one contraction function f(X,Y), e.g. four functions are given for the so-
called Barnsley fern, which leads to 24 numbers all together.
Starting the chaos game means to choose any point on the plane and using one of the
given functions after the other at random with marking all resulting points. Excluding the first
starting rules
points, where the system has to balance itself out first, will lead to the final IFS-picture. This
final form, picture, of using the transformation-rules ‘f1 to fn’ is called the attractor of the
system. For one and the same setup of an IFS there exists only one attractor, which means it
will always lead to the same final picture. If we think about the important influence of the
planning stage or the process used to develop a design for painting or architecture the IFS
underlines the same large effect of starting rules on the resulting end product.[15]
The transformation f(X,Y) given above can also be expressed by a polar transformation as follows:
transformation functions
√(A2+C2), a1=arccos(A/(√
R1=√ √(A2+C2));
√(B2+D2), a2=arccos(B/(√
R2=√ √(B2+D2));
Therefore the new set of equations is:

X(n+1)=(R1*cosa
a1)*X(n)-(R2*sina
a2)*Y(n)+E
Y(n+1)=(R1*sina
a1)*X(n)+(S*cosa
a2)*Y(n)+F.
‘R1’ and ‘R2’ are the scaling factors, ‘a

a1’ and ‘a
a2’ two rotations and ‘E’ and ‘F’ two translations.
Some sets of functions and the correspondent values, e.g. for a fern, can be found in
‘Scientific American’[16]. Representing the same fern in good quality on a television screen
compression
requires defining and fixing over a hundred thousand points. This means that if we find the set
of functions, the algorithm, of objects we can reduce their quantity of information – this is
important in the case of compressing pictures on the computer. How can we get the set of rules
for any picture? We choose certain transformations, by trying out on the screen, and apply them
to the original image. If they change the image only a little bit, then the resulting image formed
by the transformations will be similar to the original one – we have found the right rules.

3.4 DLA Models – Diffusion-Limited Aggregation Model
In physics and chemistry diffusion means a certain behavior of two different gases or
diffusion
liquids, which get in touch with each other. This behavior is characterized by the circumstance
that two different gases or liquids are mixed when they are brought together. This mixture
happens because of molecular heat emission, but the way of diffusion, its ‘form’, cannot be
forecast, which means that it cannot be calculated by mathematics.
How can diffusion then be described? Through simulation: a certain point is

marked, for example in the middle of the computer-screen. This point is the starting
point of the diffusion. Now another point, anywhere on the screen starts its
description
wanderings. The way it follows is random, analogous to the Brown's movement.[17]
The movement is stopped when it touches the stationary point, because at that
moment it is also turned into a fixed point. We can also think of a screen full of dead
cells – one cell of which is turned to life in the middle part of the screen. Touching this
living point means that the moving cell is also turned to life. The random movement
is repeated for the next cell or point anywhere on the screen – see picture 18.
The developing object looks like a map of bigger streets with some dead end roads. There
are two extreme cases for possible starting points: first, the point lies in between those streets
which means that the moving point would quickly be fixed at the nearest street and therefore
the random walk would be very short. Second, the starting point could also lie on an already
limits
living cell, which means that no random walk is possible. In both cases the random possibilities
of the walk are highly restricted, therefore the only limit of the simulation is that the new point
must not be chosen inside the growing object. In a computer simulation program this can be
avoided by drawing a circle with the largest extension of the object. The new starting points
then have to be chosen outside this circle.[18]
3.5 L-Systems – Lindenmayer-Systems
symbols
Lindenmayer-Systems, are like the IFS very close to naturally looking objects. The biologist
Aristid Lindenmayer developed this variant to describe plant-forms. Similarly to the
transformation rules of IFS, which makes ‘n’ new elements out of one by a certain rule, the
Lindenmayer-System in its first step uses symbols as a set of rules. To illustrate this stage, a
set of transformations may for example be given by: ‘l’ is replaced by ‘l + l - - l + l’.
That means starting with ‘l’, the first iteration for this example runs as follows ‘l + l - - l + l’,
the second one ‘(l + l - - l + l) + (l + l - - l + l) - - (l + l - - l + l) + (l + l - - l + l)’ and so on. But this
translation
does not lead to any picture – e.g. plant-form. Therefore a second step is required, which
translates the symbols into drawing rules. ‘l’ for example symbolizes a piece of a straight line
forward, ‘-’ an angle of e.g. 60 degrees to the right and ‘+’ an angle of 60 degrees to the left.
By that it is possible to generate self-similar mathematical fractals but also plants, bushes
and trees – see picture 19.
An example for a bush is given in ‘Scientific American’[19] with ‘l’ being turned into the
limits
expression ‘l [ + l ] l [ - l] l’. There ‘l’ once more symbolizes a piece of a straight line
forward, ‘-’ a certain angle to the right and ‘+’ one to the left. The new symbol of ‘[‘ defines

the start and ‘]’ the end of a new branch – see picture 19. The Lindenmayer-System
reduces the high information of plants through short instructions, which leads to one
possibility of application, the usage in computer-generated film sequences, where computer
power has to be reduced.
3.6 Midpoint Displacement Method

The mathematical, non-linear and linear fractals presented above are deterministic, which
means that repeating the transformations under the same starting conditions will always result
in the same figures. The midpoint displacement method, however, belongs to the category of
random fractals, such as the fractals generated by the DLA-method, which in general produce
more nature-like ‘random’ objects.
How does the Midpoint displacement method works? We draw a triangle on the screen and
mark the centers of the three borderlines. Then we move these points perpendicularly to the
construction rules
lines up or down by a random factor. The resulting object consists of four smaller triangles, one
of them is the combination of the three newly constructed points – see picture 20. The same
is applied to the new triangles, and so on. For the decision whether the center point is moved
up or down we use a coin. If the values of the displacement, the vertical intervals, decrease
slowly, the resulting object will look like a ‘young’ highly rugged mountain. In contrast to that,
any fast decreasing of the displacement factor after only a few iterations produces a smooth
mountain.
3.7 Strange Attractors
phase space
The long-term behavior of a system can be represented in the so-called n-dimensional
‘phase space’. There the specific time development is described through a trajectory, through
which the history of the system is recorded. Attraction areas, to which these trajectories aim,
are called attractors. Put into other words an attractor is a preferred position for the system to
which it evolves no matter what the starting position is. Once such a position is reached it will
then stay on the attractor in the absence of other factors.
The existence of an attractor[20] in general means for a scientific process that it possesses
the characteristic either to run in a stable, periodical or quasi-periodical way. ‘Stable’ means
attractors
that the system aims at a certain end condition, called point attractor or the fixed point of the
system. On the other hand a process is ‘periodical’ if it repeats itself through a certain interval
of time. Finally ‘quasi-periodical’ means that it lasts some time at the beginning until it turns
into a periodical behavior – see picture 21.
Another category of attractors is called strange attractors, whose name arises from their
strange attractors
strange characteristics. They consist of an infinite sequence and offer an unpredictable chaotic
behavior, but nevertheless in the phase space they occupy a sub-room of lower dimension.
Looking at neighboring trajectories their expansions follow completely different directions.
From that follows that though the system evolves to and remains on the attractor, it is not
possible to give a long-term behavior – see picture 21. This category is often applied to
represent chaotic systems.[21]

4 Dimensions ‘Fractals and Fractal Architecture’
4 Dimensions
Euclidean geometry
4.1 Euclid
Up to these days we have been used to think and talk in the words of traditional Euclidean
geometry.[01] But many complex objects described and composed by single Euclidean
sections[02] do not really reflect the characteristic of the whole real-world object, clouds and
mountains respectively do not correspond to simple geometric rules.
4.1.1 Euclidean and Fractal Geometry
co-ordinates
Fractal curves consist of infinite elements which are infinitely small and which
are, because of that, not tangible. These infinite elements are the reason why the
length increases to infinity at an infinitely small scale and by that makes it
impossible to define a point of a fractal curve by co-ordinates or describe its
position on the curve exactly after all.[03]
That is the main difference to Euclidean geometry: looking at any point of an Euclidean
curve, its position can be described by only one parameter, e.g. the distance from a starting
point. Whereas if we define a point on a coastline by its distance from another specific starting
point on the line

point, we will see that the position of the resulting point strongly depends on the scale of map
we are using for measurement. Measuring a specific distance, e.g. given in kilometers, from a
starting point on a map with a scale of 1:50.000 will lead to a new point more distant from the
starting position than measuring the same distance by walking along the coastline, which would
mean using a scale of 1:1.[04] From this follows that a coastline has to be more than an
Euclidean one-dimensional line, but it can not be a two-dimensional object either because it
does not fill the entire plane.
4.1.2 Different Dimensions

E-dimensional
The concept of dimension used in school mostly deals with Euclidean geometry. In short, in
an E-dimensional system of co-ordinates at least E-co-ordinates are needed for defining the
position of a point. Consequently a point corresponds to a zero-dimensional system of co-
ordinates, points on a line to a one-dimensional, a line on a plane to a two-dimensional and
finally a plane on a cube to a three dimensional system of co-ordinates – see picture 22.[05]
The ‘topological dimension’, however, proceeds from the fact that each structure can be
topological-dimension
reduced to a set of points. In this connection the disconnected set of points corresponds to the
dimension zero. The dimension is then a rate of how many elements of the set of points are
necessary for breaking the set: e.g. a line has the dimension of one because only one point
has to be taken out for breaking the line into two pieces. This also means that fractal curves
are still curves because intuitively the standard-arch is a connected set that can be separated
– turned into an unconnected set – by taking out only one point. In the same sense the
topological dimension of the Cantor Set is zero because there are not any two points, which
are connected – to separate them no point has to be removed.

4.2 Fractal Dimension

The following pages give a summary of the characteristics and explanations of fractal
dimension.[06] Visually the fractal dimension is the expression of the degree of roughness,
explanation
which means how much texture an object has.[07] It also shows how fast the length of a fractal
increases from one iteration to the next. Fractal dimension is not an integer in contrast to the
dimension in Euclidean geometry. The complex forms of clouds, blood vessels, coastlines or
mountains seem to have an unrestricted complexity, but they nevertheless have a geometric
regularity, their scale-independence. That means, if we analyze the structure on different
scales, we will always find the same basic elements. Fractal dimension also expresses the
connection between these different scales.
There are different kinds of measurement-methods for fractal dimension, some of which will
fractal dimensions
be explained on the next pages: e.g. for ‘true’ mathematical fractals the so-called self-similarity
dimension ‘Ds’ can be measured by the increase of length from one iteration to the next. The
dimension of coastlines and borders as examples of fractal curves can be measured by the
structured walk-method. Finally, the box-counting method is suitable for measuring the
dimension of elevations of buildings, mountains and other objects. This latter method is then
used in chapter ‘9
9 Statistics’ for the analysis and comparison of buildings.
4.2.1 Self-Similarity Dimension ‘Ds’
The self-similarity dimension ‘Ds’ is equivalent to Mandelbrot's fractal dimension ‘D’. It
self-similarity
proceeds from the fact that in a self-similarity construction there exists a relationship between
the scaling factor and the number of smaller pieces that the original construction is divided
into.[08] This is true for fractal and non-fractal structures: e.g. a line, as an example for a non-
fractal structure, can be divided into three identical parts. In this case the number of new
pieces ‘a’ is three and the reduction factor ‘s’ is one third, see picture 23. The dimension ‘Ds’
results from the equation:
a=1/(sDs)=(1/s)Ds a ... number of pieces

s ... reduction factor
Ds ... fractal dimension
line and square
3=1/((1/3)Ds)=(3)Ds
3=31 means Ds=1 in the case of a line
Equally a square can be divided into four pieces by using the reducing factor of one half. If
the reducing factor is one third then this results in nine similar parts and so on:
Ds=log(a)/log(1/s)=log(4)/log(2)=log(9)/log(3)=2
In general an object is a non-fractal Euclidean structure if there is no growth in length, area

fractal curves
or volume as one observes it more closely, which also leads to the conclusion that Euclidean
objects always have an integer dimension equal to its topological dimension. In contrast to that,
fractal curves are objects that’s fractal dimension is greater than its topological dimension. E.g.
the self-similarity dimension ‘Ds’ of the topological one-dimensional Peano curve is generated by
Ds=log(9)/log(1/3)=2. In this case the curve fills the surface it is lying on completely but if only
one point is taken off, the curve nevertheless falls apart into two pieces.

The self-similarity dimension of the Koch curve:

initiator a=1/(sDs)=(1/s)Ds
1l a=4; s=1/3;
Koch curve
1/3 generator
l
1/3
l Ds=log(a)/log(1/s)=log(4)/log(3)=1.26186
1/3l 1/3l
The dimension can also be measured by using two different scales – scale comparison:
Ds=log(a2/a1)/log[(1/s2)/(1/s1)]=log(a2/a1)/log(s1/s2)=log(16/4)/log(9/3)=1.26
comparison
a1 ... number of pieces of the 1st scale=4 a2 ... number of pieces of the 2nd scale=16
st
s1 ... reduction factor of the 1 scale=1/3 s2 ... reduction factor of the 2nd scale=1/9
Ds ... fractal dimension measured by the comparison of the 1st and 2nd iteration of the
Koch curve.
4.2.2 Measured Dimension ‘d’ – Structured Walk Method
0/0 The measured dimension ‘d’ is related to Mandelbrot's fractal dimension ‘D’ by
r0
the equation D=1+d. The illustration on the left side displays a coastline, a mountain
ridge or a border of a city on a specific scale, with the dotted line representing the
border. The dimension of this curve can be calculated by using the structured walk
expanation
method. Thereby the top of the line defines the starting point ‘0/0’ from where the
initial distance ‘r0’ is marked on the curve. Where it hits the border, the distance ‘r0’
is again marked on the curve and so on. When the end of the line or the starting
point is reached, the total length for the specific ‘rn’ is measured by using the
equation Ln=Nn*rn, with ‘Ln’ being the total length and ‘Nn’ being the number of
pieces of ‘rn’. Then the measurement is repeated by starting again at the point ‘0/0’
but this time using ‘r1’ which is half the size of ‘r0’. After counting the pieces ‘N1’, the
dimension of the border can finally be calculated as the comparison of both new
polygons, see picture 24:
r0
0/0
r1
0/0 L1/L0=(r0/r1)D; D=(log(L1/L0)/(log(r0/r1))
r0=1st distance=1stscale; L0=N0*r0=1st length;

N0= sum of r0-pieces;
formula
r1 = r0/2; L1 = N1*r1; r2 = r1/2; L2 = N2*r2;
Generally spoken, the measured length increases when the scale ‘rn’ decreases. If the
value chosen for ‘r0’ is too big, the measured dimension changes very much in relation to the
measured dimension, for example between the scale of ‘r4’ and ‘r5’. That means that the
starting length of r0 has to be chosen very carefully.[08]

Hausdorff-Besicovitch dimension: HB=log(N)/log(1/s)

HB ... Hausdorff-Besicovitch dimension
HB-dimension
N ... number of pieces by which the first line, initiator, is replaced
1/s ... reduction factor, that is the length of the parts of the generator in relation
to the initiator;
For the Koch curve this results in: log(4)/log(3)=1.2618595;
4.2.3 Box-Counting Dimension ‘Db’
different methods
The box-counting dimension ‘Db’ is equivalent to Mandelbrot's fractal dimension ‘D’.
Generally no clearly repeating self-similar structure as in the Koch curve can be found in any
real-world object, which means that these less regular shapes cannot be divided into equal
parts and by that the self-similarity dimension method will not work. But there is also a problem
when measuring the dimension of a coastline or part of it with the help of the measuring-
method if surrounding islands are to be included.
In such cases the so-called box-counting method, which is often used by calculating
the image
programs for the computer, takes remedial measures. It works very well for images prepared
as a black and white image of any object. At the same time it is important to remember that it
is the dimension of the image, which is measured, and not the object itself.
How does the box-counting method work? First a grid is put over the image, e.g. a border line
or an elevation of a building. Then the boxes, which contain part of this border or edges of the
elevation, are marked.[09] The number of boxes of the bottom row of the grid gives the unite-size,
explanation
the scale. At the next step a lower grid is chosen and again those boxes, which contain a relevant
part of the image, are marked. That means by using the box-counting method not the line is
divided into parts but the grid, which is put over the line. For calculation the occupied boxes of
each grid-size are counted. Finally, as with the measuring-methods above, different scales are
compared, see picture 25. The formula for calculating the fractal dimension is the following:
Db=[log(N(s2))-log(N(s1))]/[log(1/s2)-log(1/s1)]
N ... is the number of boxes in each box-grid which contains part of the structure;
formula
1/s ... number of boxes across the bottom of the grid – unity-size;
E.g. the box-counting dimension of the coastline of Britain measured in picture 25 between
the scale of 1/s2 and 1/s3 amounts to Db=1.31. The calculated measured dimension ‘d’ for the
scale between the unit length of 100 km and 50 km is 0.308, that is Db=1.308 .
Graphically the box-counting dimension is calculated by transforming the results, marked

boxes and the unit-sizes, into a log-log graph. The gradient of the resulting line of the log-
log graph is the fractal dimension of the image. But as I will mention later in chapter
grafical
‘7
7 Problems with Measuring’ there are some problems to cope with when using programs,
which generate the dimension. Mostly the problems result in a non-straight graph-line. That
means that there may be some shortcomings in the original box-sizes of the grids or the
quality of the image. Therefore a replacement-line has to be calculated which runs straight.
This line is the average of all dimensions given by the log-log graph.

4.3 Conclusion
The three different dimensions dealt with above – ‘Ds’, ‘d’, ‘Db’ – may produce different
different measurement methods

values for the dimension of a curve. The reasons are:
-) Self-similar structures on all scales are only found in truly mathematical fractals.
-) Overlapping curves can only be measured by self-similar dimension, because the other
two methods cannot include ‘overlapping’ parts of the curve, which means that the
resulting dimension would be lower than the self-similar dimension ‘Ds’.
-) When the box-counting dimension method is applied to a design such as the Islamic
garden layout, e.g. a grid, which is not a mathematical fractal, we will nevertheless get
a non-integer value for the dimension. This is true for a certain range of box-sizes, but
at a smaller scale the dimension will be reduced to 1. This indicates that beyond a
certain range of scale the design shows a self-similar characteristic, but looking more
closely the structure is reduced to lots of straight lines. This illustrates the range of scale
beyond which self-similarity is present.[10]
In the journal ‘Scientific American’[11] – Jürgens, Peitgen and Saupe used metaphorical
comparison with languages in their article in order to explain fractal geometry and their
characteristic elements. In the Western languages we know a limited number of letters of a
finite alphabet, which have to be put together for a special meaning. The Chinese language
consists of as many signs as concepts so that we can think of infinite elements in contrast to
language
Western languages. Each of these characters has its own meaning. The traditional Euclidean
geometry can be compared with the Western languages, where letters correspond to simple
basic objects such as the line and the circle. Complex forms are then produced by bringing
together these simple limited elements, and only after doing so, the object gets its meaning,
whereas fractal geometry, corresponding to the Chinese language in our example, consists of
infinite elements, signs or, in the case of fractals, algorithms or procedural rules, so that the
meaning is found in each part of the object.
Long-lasting processes, called evolution, produce nature and its objects. This implies that
there are many factors that influence an object in its development. The importance of these random
influences varies from scale to scale, which may lead to different dimensions. Combining
different rules, defined as affine transformations, like scaling, translation or rotation, can
produce such more natural-looking patterns. If these combinations moreover are produced at
random, the effects are even more natural, which means that random is an important factor to
get natural-looking fractal patterns.

5 Fractals and Architecture ‘Fractals and Fractal Architecture’
5 Fractals and Architecture

If architecture stands for continuing the development from the protecting caves over the
fallen down tree as a first shelter to buildings made of timber or stones and up to modern
interpretations of nature like Frank Lloyd Wright’s examples, then architecture, natural
nature and architecture

materials, time and the structure of nature may still be an unity. In this way, nature, as we
have seen in the sections before, offers characteristics of fractal geometry rather than such
of Euclidean geometry. So maybe because of similarities between nature and architecture,
with regard to material and structure, some of the fractal attributes can also be found in
buildings, in their elevations and ground floors. Thinking of self-similarity for example this is
not a new aspect in architecture, as there are similar forms on different scales e.g. in the
Gothic style. But also Frank Lloyd Wright used variations of form on different scales as a
concept of his buildings and he did not copy nature as it is offered in trees, but was looking
for the underlying structure of their organization[01]: ‘Quite a different form may serve for
another, but from one basic idea all the formal elements of design are in each case derived
and held well together in scale and character’[02]. The true mathematical fractals offer such
ideas where complex forms have simple underlying algorithms.
5.1 Visual Perception and Fractal Range

Architectural composition is concerned with the progression of interesting forms from the
zooming
distant view of the facade to the intimate details. As one approaches and enters a building,
there should be another smaller scale, interesting detail that expresses the overall intent of the
composition, which is the fractal conception. Thus fractal geometry is the formal study of this
progression of self-similar detail from large to small scales.
5.1.1 Changing of Dimensions
Benoit Mandelbrot used a wool-skein to show the changing of mathematical constancy – the
changing of the ‘effective’ dimensions. For a viewer who is far enough away from the wool-skein,
which may have a diameter of 10 centimeters and consists of a 1-millimeter yarn, it looks like a
point, a zero-dimensional object. In a 10 centimeter resolution the skein is a three-dimensional
wool-skein
object, in a 10 millimeter resolution it turns out to be a confusion of one-dimensional twines, at a

0.1 millimeter resolution each twine is experienced as a three-dimensional column and at a 0.01
millimeter resolution each column turns again into one-dimensional fibers. If we zoom in deeper
on the resolution of an atom, the wool-skein is presented as an object of finite numbers of atom-
like points and therefore it is again zero-dimensional – remember the Cantor Set. So the wool-
skein shows a sequence of different effective dimensions. Some badly defined change-over
between zones of well defined dimensions are interpreted as fractal zones where the fractal
dimension is higher than the topological one.[03]
Benoit Mandelbrot also wrote about the surface of timber being spongy but nevertheless
the beam is said to be even. This is because of the scale, in a certain ratio of size the regular
timber
and continual aspect can describe an object in a correct way. He compared this scale with
a tinfoil that is put over a sponge, which follows the surface but does not show the many
little complex details, see picture 26.[04]

5.1.2 The Scale-Range of a Wall
one-, two- and three dimensional

Observing a straight wall, built of stones of a certain thickness, height and length from a
great distance, the object may look like a straight line – e.g. having a look at the Great Wall of
China from the universe it will be identified as a long curve with no width and no height. But
coming closer the curve turns into a two-dimensional plate with a certain height and length but
no width. Finally standing on the wall it will turn out to be a three-dimensional object. Then
looking at a small stone as part of the wall a couple of meters away, this stone may be
experienced as a zero-dimensional point. But coming closer once more, this stone turns into
a three-dimensional object. This changing of dimension can be repeated down to the scale of
the atom, which offers a physical border, see picture 27.
5.1.3 The Scale-Range of an Elevation
The changing of the ‘effective’ dimension is also true for other man made objects. E.g.
the building from outside

observing a building from far away it may seem to be a point or, thinking of rows of houses,
maybe a line. Coming closer the two-dimensional outline can be recognized but no deepness
of the elevation. Windows seem to be points and lines. When analyzing the elevation from the
front it cannot even be said if it is only a stage-prop. Coming closer up to a couple of meters
away from the building, more and more details like cuts of the windows and the jumping
forward and back of elements of the facade come to one’s attention. On this scale the house
is experienced as a three-dimensional object. At some point even the roughness and structure
of the facade can be felt; it, however, depends on the material of the surface of the building at
which distance this will be. In front of the door some more details like the doorknob can be
found, which may have looked like a straight line before.[05]
Using Euclidean vocabulary implies a certain abstraction such as a plastering facade is
the facade is fractal

actually rugged but is mostly called smooth. This abstraction is taken from a certain scale, a
certain measuring: the man and the human body. From a certain distance the facade really
looks like the Euclidean flat plane, though thinking of a brick wall, there is always a structure
to be observed. This leads to the conclusion that the cut out windows and doors, beside other
details such as the structures of walls and door knobs, are what the facade differ from
Euclidean shapes at this scale, which is also true for smooth concrete facades. Therefore the
straight smooth modern style buildings are also fractal forms on closer observation.
At the next stage, when entering the house, the next level of the object is examined. There
is a difference between smooth rectangular buildings and a building of differentiated ground
from inside
plans and elevations. It is easy to take up the first type at once and explain its rooms with a
few words. In contrast to that, a ‘differentiated’ house with different heights, different materials
has more underlying information to offer. But if there is too much that we cannot understand
in the context this can lead to an oversupply of information like in a language we do not
understand. In this case the complexity turns into confusion.
5.1.4 Conclusion
“Architectural composition is concerned with the progression of interesting

forms from the distant view of the facade to the intimate detail. This progression is
necessary to maintain interest”.[06]

The principle of stepping forward of details by zooming, which is mostly found in nature,
should be taken up as a reflection of the intention of the composition from the whole to the
distant view to the detail

detail. This means that for example a round surface asks for round floor plans, which also
means that the resulting concept is understandable from the whole to the detail.
This is the fractal concept: self-similarity from scale to scale, but of course in a more general
view as with true mathematical fractals. By that, each distance, scale offers a certain attention
to details. Fractal structure therefore stands for the continuing of an architectural composition
from distance, the outer elevation, the interesting detail from outside to inside and the shaping
of the interior, which means entering and ‘using’ the respective scale.[07]
The scaling range of an object, and depending on that the grid-size for the box-counting
method, is related to the nature of visual perception. Very fine details can be observed only
distance and attention

within a range of two degrees from the center of the inspected object. But significant details
are also realized from an angle of 10, 15 and 20 degrees – see picture 28. Each scale has
a certain distance to the observed object, which can be expressed by the equation: the
distance from the building multiplied by tangent of the angle is equal to the measuring unit
size. But there is a difference whether one concentrates on a certain detail or on the whole
from the same distance. In the first case, looking at the detail, e.g. a short line in an
abstract painting or the doorknob of the entrance, objects around it will also come to our
attention. If we concentrate on a larger detail or the whole, e.g. the elevation of the
building, smaller details run out of perception.[08]
5.2 Topology
In mathematics there is one branch that analyzes form, which is called topology – in
architecture we also know this branch of topology that deals with forms. Topology[09],
which is also-called the geometry of the position, for example says that all pots with two
handles have the same form, because under the presumption of infinite flexibility and
the fractal form

compression they can be formed continually into each other without producing new holes
or closing them. It also says that all coastlines have the same form because they are
topologically identical with a circle – both have a topological dimension of one. Adding
islands off the coastline only means that the new object is topologically identical with
many circles.
Benoit Mandelbrot looked for other aspects of form in nature than this topological one.
He pointed out that though topology does not differentiate between different coastlines,
they nevertheless have different fractal dimensions. These different dimensions express
the differences in a non-topological aspect of the form, which he calls the fractal form.[10]
5.2.1 Information and Memory

memory
From a physiological point of view man can take up a total set of information of 109 bit per
second out of the rich perception offered. 102 can be dealt with in our consciousness, but only
one bit per second is saved in our memory![11]
So what does that mean to the elevation of a building? It means that a lot of information
is taken up by observing a building, of which only a little part is dealt with in our

consciousness, where it is compared with information we have already in our memory. E.g.
if one does not know what a church ‘normally’ looks like, that is if someone has never seen
knowledge of a church
any, or even does not know what it is for, the form including signs like the cross is an
absolutely new information for this person. Almost all people from our cultural background
already have an idea of a ‘typical’ appearance of a church, which may include the cross as
a sign, a church-tower, a higher middle section, rose-windows, certain proportions and
positions of windows, through which even more modern interpretations will be recognized
as a church building. That means we have seen and learned about different form-types of
churches that are in some way similar, which also means that we will have information about
details of the elevation in our memory.
The more we generally know, the more we can process certain details. That is because
the first impression confirms our picture in our memory and so we can concentrate on the
order and surprise of forms

second impression, on a smaller scale. If it is not like that, we feel insecure; the church has
a lot of new information for us to be dealt with. This is then quite similar to information we
get in a language we do not understand. But on the contrary if a church only consists of the
most typical parts of a simple reduced ‘church-form’ we do not learn anything new, we do
not get any information. The part in between is interesting for us, including what we learn if
we come closer to the building, that is if there is any new information coming up or not. The
two extreme cases of too much and too low information is quite similar to the two extremes
in chaos-theory between total confusion and determinism, but also to the complexity
generated by simple algorithms of fractal geometry and fractal dimension as the
measurement of a mixture of order and surprise.
5.3 The Box-Counting Dimension and Architecture

Throughout the following sections some fractal dimensions of elevations or ground plans
distance and its information

are calculated for analysis and comparison. The measurement method used for these
calculations is the so-called box counting method as described in chapter ‘4.2.3 Box-Counting
Dimension ‘Db’’. For better comparison of these fractal dimensions of buildings, it is important
to give the scale of the map or plan that is used, which can be interpreted as the distance e.g.
to the elevation. As mentioned before in chapter ‘5.1.3 The Scale-Range of an Elevation’ the
distance again influences the dimension through the recognizability of details – the smaller the
scale of the plan the more details can be offered. Because in general only those boxes are
counted in which some ‘interesting’ details can be found, which may be window-frames and
edges of window-strips, doors, edges of walls, the structure of the facade, but even the
surroundings or, on lower scales, door knobs, the importance of certain scales and its
information can be measured and compared.
elevation – order & surprise
As mentioned before the fractal dimension is the identification or better to say the
quantifiable measure of the mixture of order and surprise, which also holds true in design
and architecture. So an extremely high dimension would indicate that the elements,
components of the composition do not correspond to each other, which means there is a
high degree of surprise. The reason for that may be that they do not stand in a certain
relation to each other and to the whole, which means that the viewer cannot identify the
overall concept or its parts. But as the chapter ‘9
9 Statistics’ will show, this is not true for
all measured objects. There are examples where high dimensions result from the

roughness of the surface, e.g. from timber-paneling which nevertheless may have a high
order in their distribution and the elevation may display high symmetry with regard to
rhythmical distribution of windows, which means high order and simple information of the
surface. This example leads to the conclusion that the cuts in the facade like windows and
doors are not the only parts that influence the dimension and that they are not of main
importance for the structure. From this follows that high dimensions result rather from the
roughness of the facade and by that from the ‘character’ of the used material than from
order by a symmetrical distribution of elements, see picture 29.
5.4 Symmetry
Architects have used symmetry in the form of bilateral symmetry, where one side of an
image is a mirror of the other, in elevations but also in ground plans of buildings for centuries.
symmetry and rhythm

In nature such a bilateral symmetry can also be observed, e.g. in the human body. But also
symmetry developed into rhythm can be found, which means that an image is repeated along
a line, as it is true for the fingers, or that it is rotated around a point – see picture 30.[12] But
on closer observation it turns out that the symmetry and rhythm is not that strict in nature and
that there exists complex diversity – as it is also true for architecture. By that every object is
different from the other, which means for example that not all human beings are the same size
and shape, and even both sides of the bilateral symmetry need not really be identical, that is
the right and left side of the human body may have little differences. Self-similarity as an
important characteristic of fractal geometry describes nature’s underlying diversity and shows
invariance with respect to size as a new kind of symmetry.[13]
5.4.1 Symmetry and Mutation
At the end of the 19th century William Bateson developed the theory of symmetry and of
discontinuous variation. It says that where information is lost or mutated, growth returns to
simple symmetry. As an example Bateson analyzed possible mutations, e.g. that of additional
thumbs at a hand, and he found out that the abnormity has a higher measure of symmetry in
mutation
the form of an additional plane as a mirror between the new and the original thumb. This
means that the ‘normal’ thumb, which is asymmetric to the four fingers, is replaced by a
symmetric part of another four fingers, see picture 31.
In general mutation may mean additional information with regard to higher

symmetry and reduction of heterogeneity but also, as Bateson thought, that the
reduction of asymmetry and the growing homogeneousness was the result of a
reduction of information.[14]
typology or an alternative
According to Bateson, genes regulate the simple bilateral symmetries by adding

heterogeneity and differentiation as a form of organization – information hinders symmetry, as
information can be a reaction to genetic or environmental disturbance. Therefore a breach of
symmetry does not mean a loss but an increase in organization in an open, flexible and
adaptable system. In this sense symmetry is a lack of organization resulting from a lack of
interaction with stronger external forces and surroundings. But in this sense typologies are
then reduced, empty and false: Organisms cannot be related to one idealized reduced type,
they are the result of dynamic non-linear interactions of internal symmetries with changes of a

disorganized context. An alternative to typologies may therefore be an internal system of

indeterminate growth differentiated by general and forecast able outer influences, which
results in unforeseeable, dynamic and new organizations.[15]
5.4.1.a Cardiff Opera
In ‘Arch+ number 128’[16] the question of symmetry is discussed in connection with a

competition for the opera in Cardiff in an article by Greg Lynn, written in addition to his
contribution. The competition had two main tasks: first the project had to get a symmetrical
horseshoe hall and second there should be a connection to the surroundings, the historical
harbor basin. This meant that on the one hand the resulting opera should correspond to the
Greg Lynn
environment and on the other hand it should have a new identity. The interesting aspect
brought up by Greg Lynn was the thought of standardization of the surroundings with harbor
and basin and the new opera through processes of differentiation rather than simplification. It
was an attempt of a transformation of existing information into the new context. Mutation and
differentiated growth was the key to the sketch. The idea behind the concept was that the
resulting object should not be determined from the beginning, and therefore should not be
reduced to the internal rules of the object itself and external outer relations. Symmetry should
be found at each level of the sketch. In this connection theories by William Bateson from 1894
were mentioned: the loss of information goes hand in hand with the growth of symmetry, and
homogeneity is equality or lack of differentiation, see picture 32.
5.4.2 Symmetry of Fractals
Fractals are symmetrical in the one or other way: e.g. elements at different levels – scales
–, are symmetrical to one another by bilateral symmetry, translation or by combinations of
symmetry. In Bateson’s thoughts this would mean that they include less information, which may
fractals’ symmetry
be expressed by few production-rules. But nevertheless there is a high degree of differentiation
because of different scales, that is self-similarity, which, however, leads to complexity. Moreover
according to Bateson’s article true mathematical fractals do not take up outer influences – there
are first of all none be found in the computer –, which leads to systems that can only be
compared to natural forms by their underlying principles. In this way Mandelbrot introduced the
possible fractal forms of coastlines by the Koch curve, in order to show the principle of self-
similarity, complexity, generating-rules and roughness. To get more ‘nature-like’ fractals the
factor of randomness can be added or even better natural influences such as weather,
resources, availability, population, natural barriers, depending on the system, e.g. plants,
coastlines or cities.
5.4.3 Symmetry of Architecture

‘golden section’
In architecture the concept of self-similarity has very often been used in the case of
the ‘golden section’. Showing this self-similarity first a rectangle is drawn with the larger
side length being ‘a’ and the other ‘b’, with the ‘golden section’ being defined as the
relation a/b=b/(a-b). In the second step a square of the shorter length ‘b’ is put into the
rectangle. Next again a square is drawn, this time in the remaining part of the ‘a-b’-side
length and so on. The resulting image is a self-similar spiral of rectangles, produced by
a cascade of self-similar proportions, see picture 33. [17]

The rhythm, developed out of symmetry, is found in architecture in many examples like the
distribution of windows, with the distances and sizes being interrelated. But also the bilateral
symmetry with the mirroring axis in the middle front door, which has its heyday in the villas by
Palladio, is always present. Sometimes this strict use leads to contradictions as with the
symmetry at any price

oratorio of the Filippines by Francesco Borromini (1637-1640) in Rome. In this building the
main entrance is not in the middle but to the right and the outside contradicts the inside
asymmetry, see picture 34.[18]
The architect Sachio Othani, who belonged to the circle of students around Kenzo Tange,
used the other kind of symmetry, namely self-similarity in form of oblique concrete posts as the
main theme of his first larger building, the international conference-center in Kyoto. Though no
room is identical, the basic theme is always present, the oblique posts, and this helps for
orientation purposes, see picture 35.
5.5 Fractal Characteristics in the History of Architecture

Architecture as a mirror of society is also a kind of public image, which is promoted by
our time and by the culture in which we are building. The architect translates and
interprets the conscious and unconscious thoughts of society. This also means that the
architect has to face history and the present[19], with fractal geometry belonging to the
introduction
present and so it should therefore be included in the one or other way. On the following
pages the question will be dealt with if fractal geometry was present in the architecture of
the past anyway.
Benoit Mandelbrot contrasts architecture that is rich in self-similarity and posses different
scales of length – see chapter ‘5.1.2 The Scale-Range of a Wall’ –, such as Gothic and Beaux
Arts, with the curtain-walled lifelessness of modern architecture. Modern architecture, the
International Style and the Machine Aesthetic, usually have no focus, no climax above the
entrance, middle or top.[20]
5.5.1 The Time
With grand buildings of the past it often took very long time until they were finished.
Examples for such architecture are the Cathedral of Chartres, which was finished within
the factor of time in buildings
about 150 years – second half of 12th century to 13th century –, the Cathedral Notre Dame
de Paris – 1163 to 1240 – and St. Peter’s in Rome – 1506 to 1626.[21] Ideas, as in the case
of St. Peter’s in Rome, and plans of these buildings were changed and it took a couple of
restarts and changeovers during the stage of construction until the buildings were finished.
One reason was that the typical architectonical style of the time was influenced by changes
in society, political views, problems in economy and other factors. All that and the fact that
money was not always available and that the technical possibilities were limited, slowed down
the completion. In the course of time many anonymous architects were also on duty and at
work, which brought another dimension to it. The cathedral is not the result of the work of one
man but the result of the work of many men, of society and of time. Such grand architecture
needed a lot of time and changeovers until they developed into the structured interesting
buildings, which we can see today, like mountains that are formed by time. Getting a feeling
for e.g. the Cathedral of Chartres takes a day, some people even need a week, but finding all

details and getting the experience of the history of each part in connection with time takes a
life-time.[22]
5.5.1.a The Structure
The architect of today of course should not try to rebuild time development, which is not
possible either, but maybe he can regard the structure of time as a master builder. E.g. the
Cathedral of Chartres consists of sections from Early Gothic to Late Gothic, with some parts
fractal characteristics
being even older and some younger than these periods. What can be built is the character and
the structure of such a development, such as the similarity from parts to the whole, the
differentiation that nevertheless repeats the characteristics of the whole, which Charles Jencks
called the organizing depth. Naturally the cathedrals are only specific architectural tasks,
which are more or less restricted to the past, but they show some important characteristics,
such as that they are built examples of the influence of time, which brings different styles into
an overall concept, they have a rugged surface with elements that are differentiated and
repeated on different scales but nevertheless subordinate to the whole, e.g. pointed arches of
the Gothic as windows, doors, costal arch and elements of the facade, they have an overall
structure which is repeated in sections, e.g. curved waved forms in the floor plan but also in
the decoration of the Austrian Rococo, and they are ‘attractors’ for tourists and architecture-
theorists, and therefore often published.
5.5.2 Gothic
The pointed rising form of gothic cathedrals reflects the change in religion:
before Gothic times there was the imperial view of life ‘God, emperor, security’,
which then turned into the view of ‘purification, cure, uniting’.[23]
philosophy and religion

This philosophy had been translated into the pointed high cathedral that is again reflected
and assisted by the pointed arches and gables inside and outside. The period of Gothic can
be seen as the first time of international cultural exchange, because of the international
participation in the crusades. Men of different regions realized that there were similarities in
culture, but also in their national, spiritual and social problems. Therefore the new concepts
and forms of the Gothic were propagated in many countries. It was rather the time of an
attempt of feeling God – mystic – and of piety than of proving God, as it had been before. The
vertical forms of the Gothic helped to come closer to God, to get a feeling for and of God. This
shows the complexity of society in the form of its attitude to religion and architecture. Walls
became steeper and verticality was also repeated in segments and in details.
5.5.2.a One Style

world-wide similarities
In addition that the Gothic concept is found in many different variations depending on the
region, such as English-, French-, German-Gothic, the time, such as the Early-, Middle- and
Late-Gothic, different types caused by cross-section and exposure, such as basilica, basilica
with galleries, rung-basilica, hall-church, rung-hall, pseudo-basilica, gallery-hall and pilaster-
church, but also on size and usage, such as imperial or clerical. But nevertheless the elements
are similar and that can identify them as the Gothic style. This is the concept of fractal
geometry, the similarity from the whole, e.g. the Gothic as such or the pointed arch, to its parts,
e.g. sub-Gothic-styles or the pointed arch on different scales, see picture 36.[24]

There is another interesting relation between society and architectural form, namely
that Gothic was the expression of the changes in politics – the rise of France to the
greatest power of Europe by a strongly centralized kingdom. At the beginning of the
Gothic style the French king had been limited to the Ile-de-France but then the French
France
empire was united and parallel to that Gothic forms were spread around the empire.
Through analogy of time and space, the Gothic cathedral had become the visual
legitimation of royal domination – the galleries of kings on the facades of coronation
cathedrals underline this legitimation. [25]
5.5.2.b Fractal Characteristics of the Gothic
The Cathedral of Reims shows self-similarity on different scales, one characteristic of

fractals, but with respect to the ‘true’ mathematical fractals in a more abstract way. On the
self-similarity in Reims
entrance level the five arch-formations are similar to each other but they become smaller
from the middle entrance to the outer parts. The straight pointed gable is then repeated in
a smaller curved pointed arch beneath, that is then again repeated, but only in the middle
three sections. This sequence is found again on the level of the rose windows above in the
top-sections of the interspaces. The level above also offers elements that consist of a
pointed straight gable with a curved pointed arch beneath, which is then again found inside
in the middle section of the sidewalls. Looking at the middle part of the two towers beside
the walls, it offers a similar form of the interspaces of the rose window level – two small
high vertical cuts, which have pointed arches each, are collected by a pointed arch with a
straight pointed gable above, see picture 37.
cathedral in Lincolh
The tracery of the ‘Angel Choir’ of the Cathedral in Lincoln, Great Britain, also offers the
fractal concept of self-similarity. This time the lower level contains eight small but high vertical
cuts, formed by the mullions, in each case two together with a quatrefoil crowned by a pointed-
arch. In each case such a pointed-arch also holds together two of the remaining four
combinations, and finally these two resulting combinations.[26] This means that the fractal
concept holds true for three levels, see picture 38.
Now let’s have a look at the rose windows of the Gothic. The rose window of the Cathedral
of Chartres in the lateral axes – 13th century – shows a fractal concept, with the outer form
rose window
being a circle. This circle is reduced and repeated in the middle part and surrounded by twelve
smaller circles. In the outer part of the whole rose window there are again twelve circles that
are surrounded by twelve circles each, see picture 39. The window of Mary’s-chapel of the
Cathedral of Wells has another structure. In this case a three-pointed curved form is
composed similarly to the Sierpinski Gasket, in which the middle part is, however, not taken
away. The triangle-shaped object is then repeated on some different scales.
5.5.2.c Ribs and Tree-Branches

are trees self-similar
In the Gothic style the sequence of ‘single’ rooms of the forerunning Romanesque style has
been changed to a unit-room. This has been achieved by surmounting the change of main- and
side-pillar of the arcade-zone and the alternating supports respectively, which mostly run into a
six-piece vault. For the tendency of reaching great heights, the cross-ribbed-vaulting was used
inside to bring the thrust into the pillars that are supported by the buttresses and flying
buttresses outside the building. The side-support was needed to catch the shearing forces.[27]

– Gaudí improved this concept at the Sagrada Familia by bending the pillars, which made
buttresses useless. – In the Gothic style, the ribs of the cross-ribbed-vaulting meet at the
pillars, which are often a bundle in continuation of the ribs.[28] It would be interesting if there
is a connection between the thickness of the sum of parts in the pillars and the ribs, similar to
tree-branches. Leonardo da Vinci wrote that the branches of a tree after a bifurcation are all
together as thick as the branch before the bifurcation. That leads to the equation dD=d1D+d2D,
with ‘D’ being two. That also means that if the thickness of trees is taken into consideration,
then trees are not self-similar, because self-similarity demands D=D and a nearly space-filling
structure means to have a ‘D’ near the Euclidean dimension of three, see picture 40.[29]
5.5.3 Baroque and Rococo
The Baroque style shows the overall structure from the detail up to the building itself in
many examples.[30] The round forms of the architecture can be interpreted as the logical
development of the shape of a pearl, as the influence of a forerunning element. Because
of only one word, pompous or Baroque style itself, a certain impression of the building
comes up to mind, though the term style is to be used carefully. In this paper the word
style is used as a definition for typical architectural elements and not in the sense of time
Baroque
or defining a period.
So what are the characteristics of the Baroque? One characteristic surely is the strong
movement of curved forms in ground plans and elevations, the subordination of all the single
elements to the whole – this can be seen as a reference to the fractal character of this style –
broken gables and rich decoration.[31] The curved form in ground plans and elevations is also
repeated in the decoration itself, which is similar to the character of the Gothic where the
pointed arch and verticality are the underlying elements, see picture 41. The Baroque – the
time of absolute sovereigns – was the last homogeneous stylistic form. It represents the visible
universe but also all that is invisible but strongly sensitive, called the transcendence.[32]
The following Rococo was more a style of inner-decoration and less a building style,
therefore in this sense the overall concept is mostly reduced to the interior. In ‘Der große
Brockhaus – 1983 – Rococo’[33] the following entry can be found on the Rococo: ‘... Leitmotif
of the Rococo-decoration was the rocaille, an ornament that has been used pictorially in the
Rococo
grotesque-decoration and that had been developed from the asymmetric mannerism image
but also from the Baroque Italian mussel-shaped form.’ This shows on the one hand the
development of forerunner-elements that influence the following and on the other hand the
importance of a leitmotif as one important aspect for the overall concept. This leitmotif and its
variation in scale and position is maybe the reason why some buildings are more interesting
than others and seem to be more homogeneous.
5.5.4 Predecessor Buildings
5.5.4.a Chartres
‘Baedekers travel guide-book France’[34] writes about the Cathedral of Chartres: ‘After awful
fires – in 743, 858, 1020 and 1194 – of the older churches, which had eventually been built on a
Gallic-Roman holy shrine, most parts of the house of God were rebuilt between 1195 and 1220.’

– Many buildings of the past can be found which were built on the site and/or the foundations of
older ones, sometimes because of religious or political reasons and sometimes because of
economical ones. One example is the main building of Schönbrunn Palace that used parts of the
influence of the place

foundations of the predecessor building, the so-called ‘Katterburg’. The place and the building
are formed together; the predecessor building influences the new one in size and dimension of
construction. – Back again to the Cathedral of Chartres: ‘The facade had been built between
1140 and 1160 in the strong forms of the Early Gothic and was not hurt by the fire of 1194. ...
The steeple in the south, finished in 1170, is an example of the rich stylish Gothic. The north
steeple got an additional finish between 1507 and 1513. ...’ This describes the richness of
different styles within one style – strong Gothic and rich stylish Gothic –, which also means that
a couple of centuries are represented in this cathedral as in many others too. But the building is
nevertheless seen as a whole and not as a jumbled together mass of different parts.
5.5.4.b Templo Expiatorio de la Sagrada Familia in Barcelona
One of today's examples of a big cathedral-building site is that of the Templo Expiatorio de
predecessor building
la Sagrada Familia in Barcelona by Antonio Gaudí, who became supervisor of the building
works in 1883. The original plans by Villar intended a rigorous Neo Gothic cathedral whose
crypt had already been dug out and the columns for the apse had been started. That means
that Gaudí also depended on a predecessor building, and so he for example could not change
the axis of the cathedral nor could he change the number and the look of the columns, as he
would have done otherwise. Antonio Gaudí interpreted the Gothic form of a cathedral in a new
way, by cleaning the Gothic of all superfluous shapes or by further developing the forms, and
thus overcame historism, e.g. he maintained the Gothic windows but he added different
segments of a circle for balance.
The facade looks like a big sculptured book with scenes and symbols from the Bible, of
facade algorithm
Catholicism. A ‘building-algorithm’ can be expressed through the rule: show scenes of the
Bible using a certain scale and position, at which upper ones should be presented bigger. The
position e.g. can be defined by the importance of the scenes or chronologically: Gaudí
separated the scenes of Christ on earth, Christ as the liberator and Christ as the judge of life
and death on the three main facades.
The number of twelve towers is a symbol for the twelve apostles, and the upper parts of the
short description
towers look like miters as a symbol for the development of bishops out of the apostles. They
are built in the form of rotational paraboloids and not in ‘Gothic pointed forms’. Seven chapels
are placed around the altar in a fan shaped way; the altar is the center like a point attractor. In
addition to that he cleaned the altar from all decorations, which in former times, such as in
Gothic times, almost made it disappear.
Antonio Gaudí further developed the construction of buttresses and arches of the Gothic
and improved the static – he used e.g. catenarian-models and cable-models with sandbags
static and nature
that showed the reverted interrelation of forces. He did not like the huge buttresses of the
Gothic as we find them in the Cathedral of Cologne and therefore he used sloping pillars
in connection with parabolic arches. These forms look like trees in a forest, an analogy to
nature and also to fractal geometry? The column is no longer just a vertical straight
cylindrical Euclidean form but it follows the same rules as a tree that is formed by weather,
from stem to branch and to the widely ramified boughs. Antonio Gaudí not only brought

nature into ornamental forms like the designers and architects of the Art Nouveau but he
also used the construction principles of nature, its underlying structure, e.g. the
interrelations of the forces of a tree. This can be one interpretation of fractal geometry in
architecture; fractal nature is brought into architecture by construction and not only as a
decorative element on the surface. ‘Fractal form follows natural construction’, the surface
is influenced by its underlying structure, see picture 42.
Antonio Gaudí thought of a flat piece of iron, which had to be formed into a three-
dimensional object by the smith, he formed the whole out of a flat ground. Mies van der
Rohe on the other hand built with flat straight elements, which is a completely contrary way
of architectural thinking.
The Sagrada familia is built in the spirit of the Gothic with the influence and a new
interpretation prevailing at the turn of the century. Up to now it has been a building site for over
acquisition
a hundred years, mainly for the 20th century, and according to Gaudí its completion will still
last a long time period which will bring up different styles, construction-methods, materials and
forms, but all based on the underlying idea of Gaudí’s handed down models and sketches. In
this sense after the death of Antonio Gaudí in the year 1926, reinforced concrete, as a new
development, was used for the floor construction.[35]
5.5.5 Modern
In ‘Fractal Geometry in Architecture and Design’ Carl Bovill measured the fractal dimension
of a cubist painting by Le Corbusier and wanted to show by that the lack of interesting detail
from a certain scale onwards in modern paintings and buildings.
Is there any connection between purism and nature and if yes where is it to be found? Put
basic natural laws

very briefly, cubism believed that objects constructed, built and made by man are tested for
their fitness of purpose as it is done in nature. In this sense the same laws of economy and
fitness as those found in nature test these objects. The result of this is that there is a basic rule
that leads to general harmony and order. So the reduction to primary Euclidean shapes and
basic colors was an attempt to show the ‘basic natural laws’. But as Bovill’s measurements
illustrate these natural basic laws were only translated into paintings and buildings up to a
certain scale.
5.5.5.a Le Corbusier – Purism
According to Bovill[36] Le Corbusier's painting ‘Nature Morte à la Pile Dassiettes’ for

a painting’s dimension
example has a fractal dimension of 1.53 on a scale between 1/13 and 1/26 of image size,
which is quite similar to fractal dimensions found in natural forms. Looking more closely the
painting becomes less interesting and the dimension between the scale of 1/52 and 1/104 is
only 1.14 – D(26-52)=1.28 –, which means that on this scale-range there is an obvious lack of
details. This also shows us that for purism the influences of man-made machines seemed to
be much higher than those of the principles of nature. That may be the reason for the limited
range of scale, on which natural fractal dimension is presented. In contrast to that, Wright got
deeper into the structure of nature and thus brought details into his buildings, which lead to
similar fractal dimensions even up to small scales – for my own measurement of the same
painting but between different scales see picture 43.

The same phenomenon can be found at the ‘Villa Savoye’ by Le Corbusier, a so-called
a buildings’s dimension
‘machine for living’. This is an example for man-made or man-constructed geometry in contrast
to natural geometry as it is represented by fractals. The dimension of this building, measured
for a scale-range of 1/16 to 1/32, results in 1.42, that for a scale between 1/32 and 1/64 in 1.33.
This means that the building does not reach the Euclidean shape as quickly as the painting.
But looking even more closely on a scale of the window-strips, the dimension drops to 1.0 –
D(64-128)=1.17! That means that from a certain scale onwards there is a lack of detail and the
dimension turns to Euclidean geometry.[37] What is missing then is the natural, structured
depth, see picture 44.
5.5.5.b Nature and Construction
The differences between Frank Lloyd Wright and Le Corbusier can be found in their
different basic understandings of constructions. Le Corbusier used and looked for materials
natural concept
and forms from man-made machines, from industry. He believed in the strength of the ‘laws of
efficiency and fitness in nature’ and translated that concept into his buildings. Frank Lloyd
Wright on the other hand believed in the development of natural materials, the former
translation into architectural motifs, and their development up to now. By using natural
materials in a way nature does, complexity and order are produced. As the IFS-method
depends on the rule and not on the starting image, the same is true for the differences
between Le Corbusier and Frank Lloyd Wright. The rules defining how materials have to be
used, in a natural or industrial way, are important not the starting material or program.
Beside that Le Corbusier also used the idea of connecting and relating parts of the facade
by an overall idea, which has been said to be a fractal concept. The elevation of Ozenfant's
studio for example is constructed by a rule of transformation that is used for the whole facade
and even for the proportions of the windows, see picture 45.
5.5.5.c Euclidean and Fractal Concepts
Euclidean
If we call the modern style more Euclidean-like, the overall concept is meant, the
rectangular walls and the cubical forms that refer to Euclidean geometry. This concept
comes up by zooming closer to the object, because then it really turns into Euclidean forms,
the dimension turns to one.
Nature has an organizing depth and this is what really makes it fascinating from the large
to the small. In comparison to architecture this is also the concept that makes Gothic
cathedrals, rural houses and some of the so-called organic architecture so interesting and
diversified. Naturally this is only a generalization and of cause there are exceptions as some
villages with smooth rural houses – which have to be looked at in connection with their
exceptions
surroundings and with the materials available – or the church Notre-Dame-du-Haut in

Ronchomp by Le Corbusier – where walls and roofs are hollow concrete shells and the whole
building has a plastical bodily effect, which is a contrast to the rectangular Euclidean style of
the ‘other’ Le Corbusier –, but mostly this statement is true. ‘Good’ architecture lies between
the smooth Euclidean and a too complicated geometry and can be characterized as
deterministic chaos that is brought into shape by fractals. Too much complexity leads to non-
understandable architecture with very high fractal dimensions.

5.5.5.d De Stijl
De Stijl, as an example for modern architecture, demanded ‘measure and number,

clearness and order, standardization and production in quantity, perfection and best
De Stijl
realization’.[38] It was the style of interpretation and realization of the geometric abstraction
of Piet Mondrian – the concrete painting – in cubist architecture. Each surface corresponds
to a certain part of the house such as piers and windows, which are arranged in differently
deep spaces. This is the Euclidean way: differentiation of the ‘vocabulary’ of the building in
space, construction and segments.
5.5.5.e Conclusion
Overstressing the unity, the ‘overall-society’ builds for the masses only, which was also true
for the modernists who only built for the middle class without looking at the different layers of
society. In the modern style of the 1920ies, the angles of buildings had to be rectangular and
their facades had to be smooth, white painted surfaces. If we regard our homes as our third
democracy
skin, it seems that we cannot think of such house-types, because a skin is something organic,
nature-like and not a sterile, rectangular Euclidean thing. That made Gropius say we had to
wait until moss comes up and man lives in those white modern buildings, because then the
smooth surfaces turn to life.[39]
Society can be shown through infinite possible variations and combinations of parts of
society, which are then integrated into the character of buildings – as a mirror of time.
Following from that democracy is not achieved by standardization but by variation and by
showing all parts in the right combination.
5.5.6 Organic Architecture
Organic architecture as such has never really been defined, but what it mostly stands
for is the inspiration or the underlying structure that is used for construction or form of
inspiration
buildings. Materials for architectural constructions come from nature and by that are
used as architectural motifs, e.g. the Egyptian columns in form of papyrus plants, see
picture 46. In this sense forms have been developed up to now. But mostly it is rather
the concept than the form that we look for in nature. Abstract steel-’tree-twigs’ as
constructive columns, shell constructions from the inspiration by sea urchins may
therefore also belong to organic architecture.
5.5.6.a Frank Lloyd Wright
The organic architecture tried to achieve harmony between architecture and landscape.
the prairie-houses
Its origin lay in the Art Nouveau that orientated itself on the biological knowledge of the 19th
century. Frank Lloyd Wright was one of the most famous representatives of organic
architecture. Proceeding from the principle of Sullivan to develop a plan from inside to
outside, Frank Lloyd Wright built organic houses that grew out of the conditions of the
environment with regard to purpose, material and construction.[40] The most important
examples are the so-called prairie-houses where he brought nature and man-made
architecture in harmony. Some characteristics of this kind of dwellings are the low
extensive ground plans, through which the function determines size and position of rooms.

The hall, eating place and living room are no longer separated but flow into each other,
rooms rise out of the center and determine the outside, flat cubes are connected with
nature by large overlapping, gently sloping, roofs – influences from Japan –, terraces and
gardens go together, the chimneys are heavy-set and the materials timber and stone are
used. The elements of the building are strictly constructed out of the architectural function
and are not covered in any way, see picture 47.
5.5.6.b Structural Similarity
Thinking of concrete shells and similar examples there is something missing if we want
the character
to call them organic architecture. When Frank Lloyd Wright called for inspiration from
nature he did not copy nature but translated it into architecture. What he was looking for
was the importance of a simple, specific form, which would be the expression of the
building. From this basic idea and basic form the whole and all other formal elements are
derived which means that the formal elements are held together in scale and character –
complex forms are developed out of basic rules or ideas where smaller parts are the
‘mirror’ of the whole.[41]
Like in the case of fractals, which are generated by certain simple rules, these rules are
changed to a basic idea, a simple, specific, single form that runs through iterations. The
more elaborate details can be found, the more iterations are used in the building. In Robie
many algorithms
house by Frank Lloyd Wright for example the cascade of interesting detail comes down to
the stained glass designs of the windows, see picture 48. If only a few iterations of the
idea are brought into a building, the character may be the same but only up to a certain
scale – from then on it will become smooth. Beside that, as there exist different algorithms
for different types of trees, clouds and other natural forms, which means that there is not
only one rule for all organic forms, the same is true for architectural concepts, which
cannot be reduced to only one rule, either.
5.5.6.c The ‘Unity Temple’
When Frank Lloyd Wright called for the underlying structure of nature in design, he for box-counting dimension
example thought of the canopy of a wood. This is true for the idea of ‘Unity Temple’ where
he thought of the light that is filtered through a forest with a complex form in the upper
regions, the branching zone, simpler forms at the trunk zone and again more complexity at
the root stage on the ground, with the leaves lying there. The fractal box counting method
can also measure this similarity between the canopy of a wood and the ‘Unity Temple’. Thus
the fractal dimension – measured by Bovill – of the upper part turns out to be about 1.42,
that of the middle part about 1.22 and the lower part 1.67 – for all three measurements the
scale-range is defined between 7.5 and 3.7 feet.
In ‘Fractal Geometry’ [42], Carl Bovill compared the Robie house with the ‘Unity
Temple’ – see also other examples for a temple on picture 49. His result for the
comparison
dimension of the Robie house between the scale-range of 12 and 6 feet, measured with
the box-counting method, is 1.485, that of the ‘Unity Temple’, between the scale-range
7.5 and 3.7 feet, 1.482. This shows that the underlying structures are similar, but of
course the forms are certainly different.

5.5.6.d Eero Saarinen
In ‘Architekten – Begegnungen 1956-1986’[43] Helmut Borcherdt repeats parts of a

lecture of the architect Eero Saarinen in Munich of 25th October 1960. The contents of this
the environment
quotation are the planning of his own new house, which could not be built because of his
death one year later. The building-site was a densely wooded, rocky bank of a river. – From
the point of view that buildings, which take up the structure of the environment, have a
fractal dimension similar to that, this building-site would cry for a higher dimension. This is
because fractal dimension is the measurement of roughness or detail-richness. – In Eero
Saarinen’s opinion the surroundings would ask for an unconstrained, dark romantic
building – of high dimension. As Saarinen had mentioned at the beginning of his lecture
about architecture as a whole, this suggested that the romantic unconstraint element had
to be maintained in the inner parts.
In this connection he had to solve the problem that on the one hand he wanted to create
the building as a unit and on the other hand with the right transition from a wild romantic
outward appearance to a more formal interior. This is the leitmotif, but the interior design
of the rooms had not been solved. The inner parts that are characterized by Saarinen's
leitmotif
lifestyle would not be that unconstrained, which means that the overall concept had to be
transferred by another rule or, better to say, adjusted. This is quite similar to the modified
structure in other concepts of e.g. the competition for the opera in Cardiff by Greg Lynn,
where an overall concept has been modified in parts by outer influences that nevertheless
fit into the whole.
5.5.7 Contemporary
earth art architecture

One category of ‘new architecture’ is the ‘land-form’ – earth art – architecture as it is
mentioned in ‘Arch+ number 141’.[44] Under the headline of ‘computer architecture’ we find a
development of computer generated buildings that take up mathematical theories – such as
folds, fractals, chaos, complexity, algorithms – and translate them into architecture, see picture
50. ‘Land-form’ architecture stands in correlation with the theory of complexity, but also with
cosmology and non-linearity. It is also an attempt of achieving harmony between modern
architecture and the city. This kind of architecture wants to be in competition and correspondence
with man-made and with natural environment. Most of other modern architecture in cities is
monumental, crushing or uses stereotypes – architecture as articulated landscape.
5.5.7.a Zvi Hecker
Out of metaphors the Israeli architect Zvi Hecker developed a ‘land-form’ building, the
Jewish school ‘Galinski-Schule’ in Berlin. The overall geometry is taken from a sunflower, which
Galinski school
connects snake-shaped corridors, mountain-stairs and fish-shaped rooms. The sunflower is

placed in the center of the object, representing the heart of the school – a circle-shaped lawn
–, while the fish-shaped parts are grouped around. The reason why it is nevertheless not a
pressed constructed pure formalistic building lies in the two other systems: one is a rectangular
grid, as the ‘man-made part’, and the other is a concentric circle that is put over the sunflower.
This results in self-similar but not identical rooms. Through the sunflower form the architect was
able to let the existing trees around survive, which is correspondence with environment, see
picture 51. Another positive aspect is that of urbaneness, which is reached by small, curved

streets and paths, similar to a historical city, where shapes and places have been formed at
random and by history.[45]
The complexity is produced by the overlapping of different constructional systems, which

makes orientation more difficult than in simple logical systems with only one grid. To get a
better orientation in the complex form and for construction, the particular parts have been
complexity
painted differently, so the three geometric systems are pink, green and blue. That makes the
complexity of the building readable, which is often found in ‘land-form’ architecture. In many
examples of similar buildings simple materials, simple forms and abstraction produce
complexity like in the Jewish school ‘Galinski-Schule’ in Berlin. This is once more the concept
of fractal geometry: to get a complex object out of simple rules or algorithms. This may be the
difference to ‘modern-architecture’: complexity in contrast to simplicity.
5.5.7.b Greg Lynn
The contribution for the competition for the opera in Cardiff by Greg Lynn shows on the
one hand the use of self-similar, true mathematical fractals and on the other hand the
interpretation of symmetry by Bateson – see chapter ‘5.4.1 Symmetry and Mutation’. The
fractal-form was developed out of the form of the oval harbor-basin next to the building-site.
The coastline had been analyzed for self-similar patterns that were found in inlets of bays,
rivers and harbors along the coastline. This ‘catching’ of water by land was used as the
Cardiff
concept of the surroundings. The fractal generated out of that has its origin in an oval that
is copied to the right side by 45 degrees with a certain reduction-factor and to the left by 90
degrees with another reduction-factor. A new arrangement was made because of the
demands of an opera, like the connection of foyer, stage, cloakroom and shops. This system
was then modified and controlled by general information from the context, with the ovals
being rotated, tipped out and arranged in the way that they could be connected with
surrounding roads, building-walls and views. The environment mutated the fractal with
higher symmetry, according to Batesons' theory. The resulting form has new, additional
inlets of the coastline.[46]
5.5.7.c Bernard Tschumi no defined boundary
The Parc de la Villette in Paris represents an example for ‘deconstruction’. For that concept
Tschumi divided the whole program into pieces. These pieces were then placed on a grid. The
resulting structure consists of point attractors, straight galleries and random curved
promenades. These three parts together result in a complex structure that has no starting and
ending and no defined boundaries as it is true for the Julia sets, see picture 52. In general
fractals do have a beginning but no ending, because of infinity iterations – zooming into the
border of the Julia sets means to find more and more complexity.[47]
5.6 The Influence of the Surrounding

organizing depth
Our natural surroundings, our society, our life consist of differentiation but nevertheless the
whole always looks like a unity. That is because in all the parts, sections of this unity we will
nevertheless always find similarities to the whole – Charles Jencks called this phenomenon
the organizing depth.[48] In a society based on agriculture not only the economy is adjusted to

agriculture, but also lifestyle, men, buildings. This circumstance is reflected by the landscape
as if it were a mirror – it is formed by typical courses of roads, farmhouses and villages.
5.6.1 Material
Most of the indigenous buildings integrate part of nature through the available
materials. Thinking of regions with soft rocks around, man-made dwellings but also
stone
freestanding protective walls often consist of these natural stones. No matter whether
they are lying free on the surface or are roughcast, the character of the conical, rugged,
baggy walls will be similar to the surroundings. The house itself has a certain similarity
with the rugged mountains through steep roofs, massive rugged walls that are repeated
in the detail – the wall itself.
The same is true for villages and indigenous buildings in areas with large forests. Again
the character of a wood is taken up in the house itself, as in a log cabin, where the stem
is moved from a vertical to a horizontal position, or by timberwork in the balcony and in
wood
the gable. Wooden tiles are also such identification with the surrounding area. So one
source of organizing depth lies in the usage of local material, with the possible variations
and results being nearly infinite. Through the material, the parts as well as the whole
consist of a similar character and structure, it is the red thread of the object.
5.6.2 Indigenous Buildings
region
Indigenous buildings generally have a deeper scaling than modern houses. That means that
coming closer to the surface, there will always be some new smaller interesting details, similar
to the concept of the houses by Frank Lloyd Wright. The different kinds of rural houses depend
on the specific region, the available materials, the size, the usage and inhabitants of the building.
In the comparison of two houses taken from ‘Alte Bauernhäuser in den Dolomiten’[49], one
two different dimensions – two native-houses

native-house in Gosaldo and one in Borca, the differences of the two regions are visualized.
The elevation – the plan of the facade – of the house in Gosaldo seems to be smoother, which
is pointed out by the box-counting dimension of the smaller scales:
D(10-20)=(log105-log42)/(log20-log10)=1.322 for Gosaldo

D(10-20)=(log208-log66)/(log20-log10)=1.656 for Borca
The more one zooms into the plan, the lower the dimension will be, which also means that
the elevation then offers less additional information and fewer details respectively – the facade
does not seem to be very rugged. The elevation of the house in Borca on the contrary offers
more and more details the smaller the scale becomes. This is again shown in the fractal
dimensions of the surface:
D(20-40)=(log283-log105)/(log40-log20)=1.430 for Gosaldo

D(20-40)=(log688-log208)/(log40-log20)=1.726 for Borca
In the case of Borca the dimension increases from the scale-range of 10 boxes to 20 boxes
to the scale-range of 20 boxes to 40 boxes, in the example of Gosaldo it decreases. Comparing

the dimensions of the plans of both facades on the same scale shows the difference in the overall
view, the elevation of Borca has a box-counting dimension of e.g. D(40-80)=1.702, the one of
Gosaldo D(40-80)=1.461. This is caused by the different structure of the elevations: the one of
Gosaldo offers a smooth symmetrical front-facade with one balcony, symmetrically arranged
comparison
windows and a small additional section on the left side. The front-elevation of Borca consists of
an outdoor staircase, some balconies with wooden carved banisters, asymmetrically arranged
windows, timberwork on the gable, a bigger additional part on the back left side and some parts
have different heights. That points out that the box-counting dimension indicates the roughness
and differentiation of the whole building – the plan respectively. The more rugged the elevation
– picture of the facade – is the higher the dimension will be. Besides one can observe the
decrease or increase in information when zooming closer, see picture 53.
5.6.2.a Translation
Through the factor of time, a random process is brought into the building, as it is found
in nature with mountains, trees, clouds, coastlines and rivers.[50] If the clustered random
cascade in nature is fractal, then the cascade of rhythm and detail in indigenous buildings
application
may also be fractal and the measured fractal dimensions of the environment, elevation and
detail will be similar.[51] Beside that, when planning in the surrounding of any other existing
buildings, their character can be analyzed and interpreted as a sort of planning-aid. The
structure includes the widths and heights-diversity, similarity of the characteristics of
bases, middle parts and tops, window and door-types and their sizes, ledges and setbacks.
These parts can be given by a fractal-dimension range and translated into a fractal rhythm.
This ‘new’ rhythm can then be used as a planning tool for setbacks, window sizes and
distribution, construction grid and more, see picture 54.
dimension as a helping tool

5.6.3 Landscape and Building
Certain surroundings demand certain fractal dimensions of the building and by that the
fractal dimension can be used as a critical tool to find out whether certain buildings or groups
of houses fit into their environment or not. Short time ago this could not be measured. But
looking at some examples of the measurements by Carl Bovill will make us believe that
harmony between man-made objects, such as buildings, and the surroundings can be
indicated by their similar fractal dimension. The reasons often lie in the harmony of the
materials used and the factor of time.
5.6.3.a Kind of Surrounding
If the environment is very rugged it asks for a house with greater detail-richness – which is
fitting or contrast
automatically brought into native-buildings by the use of traditional, available materials. In

contrast to that, a house with a much smoother elevation may fit better in flat, uniform
surroundings. Nevertheless these two different possibilities also work the other way round. So
it is possible to use different fractal dimensions of building and environment to attract attention.
These possibilities hold true for a natural surrounding as well as for an artificial environment.
In this way the fractal rhythms can be used to produce planning grids that utilize the rhythms
of nature as a source of layout inspiration. For example the fractal dimension of a mountain
ridge behind an architectural project could be measured and used to guide the fractal
rhythms of the project design. The project design and the site background would then have

similar rhythmic characteristics. Time and local materials are important factors for this process
as well. The composition of indigenous buildings has a random cluster, which is similar to the
clustered randomness of mountains, clouds and trees.
In this sense the fractal dimension of surroundings can be used on bigger scales like
mountain-ridges, forests and typical landscapes for calculating the dimension of the site plan
scale
and overall view, or on smaller scales like single trees and immediately surrounding landscape
for elevation or floor plan of a building.
5.6.3.b Two Different Regions
There is an example in ‘Fractal Geometry’[52], where Carl Bovill compares two different
Sea Ranch and Nantucket

houses with their typical environment. One example is the sketch of the condominium building
at Sea Ranch, California. The coastline there has a self-similar structure in form of a very
rough edge from the scale of the bay to the scale of pebbles, which calls for complex shapes
of the building. In general many of the buildings in this region and especial the condominium
building by Moore – built in the sixties – offer the complexity of the coastline, which means that
they are like a mirror of the environment. The forms of the surrounding rocks of the coastline
and the building itself have similar characteristics. Nantucket, Massachusetts in contrast has
a very smooth coastline. This smoothness is also included in the texture of the waves washing
up on the beach, which calls for simple basic shapes. The natural concept is not only shown
in the forms of the coastline but also in form of the site-plan and the elevations of the
indigenous settlements, see picture 55.
5.6.3.c Including the whole Environment
similar dimensions
For showing how buildings may correspond with their surroundings, Carl Bovill analyzed
the example of a row of traditional houses in Amasya in Turkey. He found out that for this
example the fractal dimensions, determined by the box-counting method, of the
environment such as the mountain and the river, were very close to dimension of the site
layout and elevations of the buildings, see picture 56.
The cluster of traditional dwellings was established in the 3rd century B.C. and is
situated along a river at the foot of the dominant mountain ridge. The measurement of the
fractal dimension of the row, which is represented as a clustered random cascade of
rhythm in height and width, results in 1.67 for a scale-range from 6 to 3 meters and rises
the measurement
to 1.83 on a scale-range from 0.75 to 0.38 meters. The site-plan, that includes the bends
of the river, results in a fractal dimension of 1.63 for a scale-range from 50 to 25 meters
and 1.1 for 3.13 to 1.56 meters. The mountain ridge, which is reduced to a sketch of the
characteristic edges, has a fractal dimension of 1.63 for a scale-range from 100 to 50
meters and 1.4 for 25 to 12.5 meters. Carl Bovill then interprets this data material. Thus it
turns out that on a scale-range of 6 to 3 meters the site layout runs out of interesting
details, which is the same scale at which the details of the elevation of the houses becomes
the attraction.
Two important findings result from this example. First the degree of abstraction of the
mountain-ridge influences the results of measurement, as this is true for the elevation of
the building, which is represented in a certain characteristic abstraction. This leads us to

the second finding, namely that the scale of the plan has to fit to the scales that are
interpretation
measured by the box counting method. This means that the smallest boxes that are
measured have to be of a certain size and have to include the scale that is represented
on the plan.
5.6.3.d Summary
Reasons for the similarity of fractal dimensions between buildings, environment and site-
plans are the surroundings in form of available material for buildings, which may be timber,
stone, clay, but also in form of defense, demonstration of power, and the form of the
repetition
environment itself. Flat or rough mountain landscapes give the people who have been living
there for hundreds of years a specific character and behavior, in the same way as light, i.e.
daylight because of north-south distribution, weather, religion and society do. Buildings in their
function as mirrors reflect those different characters. The kind of characteristic features may,
however, be different because of outer influences, such as the contact with other cultures,
adaptations or the different importance of certain factors.
5.7 Analysis
problems
As the previous pages have illustrated, the box-counting dimension of elevations and
ground plans can be measured and then compared with each other or with its surroundings.
But there might arise some difficulties from the measurement itself, which are dealt with in
chapter ‘7
7 Problems with Measuring’.
In chapter ‘9
9 Statistics’ some of the comparisons of different house types, such as double
houses, one family houses, dwelling houses or farmhouses, under the usage of different
parameters, such as including tiles and furniture or not – concerning ground plans – or
including the surface material or not – concerning elevations – are mentioned. One question
arising from that is: whether different types of houses have different ranges of fractal
dimensions or not. Because of different scales and qualities of plans no unequivocal results
statistics
but some tendencies can be given. In this sense elevations of smaller architectural tasks seem
to have a lower average fractal dimension – around 1.48 – in comparison to bigger dwelling
houses and public buildings – around 1.55. But the dimension of the latter mentioned group
might also result from the generally smaller scale of plans used or of too little data used.
In any case, what is to be measured is the roughness or smoothness of the ground plan or
elevation. This differentiation is then worked out with the examples of farmhouses in chapter
‘9
9 Statistics’.
5.8 Helping Tool in the Stage of Planning

introduction
In the previous pages fractal geometry was mainly used as an analyzing tool in form of
calculating the fractal dimension with the box-counting method or as a means of finding out
whether buildings offer any fractal characteristics or not and if self-similarity and higher
fractal dimensions are the reasons for greater acceptance of certain styles in contrast to the
modern flat style. Now, in the following chapter, fractal geometry is used as a helping tool in
the stage of planning by getting grid-layers and possible distributions.

5.8.1 Midpoint Displacement
As mentioned in picture 20 ‘Midpoint Displacement’ the area under a parabola can be

measured by the midpoint displacement method, in which the height of a certain step is related
to the heights of the step before by the formula height(n+1)=(1/4)*height(n). Landsberg, a
mathematician, scaled the vertical lines by a certain factor ‘w’ instead of using 1/4, which is
fractal curves
expressed by the formula height(n+1)=w*height(n). To get a curve out of this formula one starts
with a horizontal line and draws a vertical line of any length from the center, connecting this
new point with the two outer points of the starting line. This procedure is repeated for the two
new outer oblique lines, this time using the formula from above for defining the length of the
displacement. The instruction is then repeated recursively until infinity. For getting more
natural looking curves the decision whether a new line moves up or down can be decided at
random. Landsberg found out that the curves produced by this formula are fractals if the value
for ‘w’ is higher or equal to 0.5 and lower or equal to 1.0. The fractal dimension is then given
by the formula D=2 – | log2* w |, see picture 57.[53]
A more complex form of the midpoint displacement method is the so-called Gaussian
midpoint displacement, in which the displacement at each step is determined by a random
Hurst exponent
choice from a Gaussian distribution. Following from that some of the displacements chosen at
random will be positive and some negative, some will be large and others small. This time the
displacement at each step will be scaled by a factor referred to as the Hurst exponent, which
varies between 0 and 1. The scaling factor ‘W’ is generated through the equation W=1/(2H),
which means that if the Hurst exponent increases the scaling factor decreases and by that the
fluctuations in the midpoint displacement process are getting smaller.[54] The fractal dimension
results from the relation D=2-H.
H.E. Hurst was a hydrologist who analyzed the variation of natural systems through time,
determining the Hurst exponent

like the variation of temperature, pressure, lake levels and tree rings, in which the so-called
Hurst exponent can compare the variation. For determining the exponent the ‘rescaled range
analysis’ is used. There the size of maximum fluctuation of a variable is observed over a range
of time scales. First a grid is placed over the curve, which should be measured, with the
horizontal axis defining the time line and the vertical grid that of the fluctuation. Now the
maximum fluctuation can be measured. In the next step the curve is divided into two parts of
1/2 of the whole time sequence. This time the measurement results in two maximum
fluctuations, of which the average maximum fluctuation is calculated. This is then repeated for
1/4 time sequences and so on.[55] Finally the Hurst exponent is measured through a log-log
graph of the maximum fluctuation range versus the time scale. In that graph the slope of the
line determines the experimental Hurst exponent.[56]
5.8.1.a Fractal Rhythms
By rescaling range analysis, the fractal dimension of fractal rhythms in nature or man made
compositions can be calculated, including the distribution of trunks in a wood, columns and
grid
walls. The main difference between Euclidean and fractal-rhythm-based grids lies in the
complexity and in the lower symmetrical distribution of the latter, in which it is valid that the
higher the fractal dimension of the grid, the more clearly the differences of the spaces, the
jumps of width, can be seen. So again the fractal dimension is a measurable characteristic of
order, with low dimensions near one, and surprise, with higher dimensions up to two.[57]

As examples for fractal rhythms, buildings by Frank Lloyd Wright are valid, such as the
fractal rhythm
grid of the major plan elements of the Willits House indicates. For measuring the fractal
dimension of the grid the rhythmic variation can be translated into a bar chart, in which the
different heights of the bars represent the different widths. Then the Hurst exponent and
from that the fractal dimension can be determined. The horizontal width of the bars does not
affect the calculation, see picture 58.
Chapter ‘5.6.3 Landscape and Building’ dealt with similar fractal dimensions of the
elevation, e.g. the window strips, and the surroundings, e.g. the site-plan and mountain ridge,
which results in a homogeneous whole. The fractal dimension of the environment can be
measured by the box-counting method or if it follows a certain rhythm by the rescaling range
analysis. The measured fractal dimension and the Hurst exponent respectively of the
translation
environment can then be used as a source for producing a fractal ‘noise curve’ by midpoint
displacement. From this fractal ‘noise curve’ a horizontal center of the distribution is drawn as
a reference-line. At the next step a grid is placed over the curve, with the grid size taken from
an architectural module size. One possible module is based on the ground size M=100
millimeters – following from that 1M=10cm, 6M=60cm, 12M=120cm.[58] By this grid the fractal
noise curve can be translated into a step function. The heights of the steps are the distances
of the rhythm of a certain fractal dimension, see picture 59.[59]
Subsequently there are still many different possibilities to choose from. First there are
different curves of the same fractal dimension; each time a curve is generated the resulting
step-function looks different, but on the whole the character is similar. That means that curves
the planner
of a high fractal dimension are rougher than curves with a lower dimension, which are
smoother, i.e. there are no high jumps. Second the interpretation of the step function belongs
to the planner, e.g. the first twenty steps may form the vertical rhythm, whereas the following
are used as the horizontal ones. But it is also up to him which lines are decided to be walls
and which open fields.
5.8.1.b Visually Distinguishable Size
The problem when translating of a noise curve into a sequence is the distinguishable
size differences between the single widths. Carl Bovill solved this problem with the
architectural scale based on human ability by Dom H. Van Der Laan. This architectural
translation
scale is based on the maximum distinguishable size differences between certain elements,
with the major whole being defined by 7 units, the minor whole by 5.25, the major part by
4, the minor part by 3, the major piece by 2.25, the minor piece by 1.75, the major element
by 1.25 units and finally the minor element by 1 unit. Through the step function a sequence
of values can be given depending on the heights of the steps, e.g. 1, 3, 5, 4, 7, 6, 3, 6, and
so on. These values are then translated by means of the architectural scale of Dom H. Van
Der Laan, in which 1 is defined as the minor element, 2 the major element, ... and finally 8
being the major whole, see picture 60.[60]
5.8.1.c Some Examples of Fractal Rhythms

applications
Smooth prefabricated concrete walls for noise abatement of the same height can be
interpreted as two-dimensional areas of nearly endless width in the sense of Euclidean
geometry. These boring unity walls in most cases have no relation to the environment such as

a forest-top situated behind them, which mostly offers a fractal dimension of 1.3, or a mountain
ridge. This fractal dimension can be used as the starting value for a fractal rhythm, generated
as a step function of a noise-curve. The resulting concrete wall then consists of prefabricated
elements of different widths and heights, which offer a mixture of order and surprise in the
same way the background nature does – the fractal rhythm leads to a variation of a complex
rhythm, where expectations are confirmed, see picture 61.[61] Fractal rhythms can also be
used for the sizes and sequences of window strips, but also as a planning grid for different
heights and widths of the front view of row houses.
In comparison with fractal rhythm sheet music, as a source of fractal distribution, can be
used for design. The notes of the music are then translated into a fractal rhythm by interpreting
the staff lines as different heights of a step function. By that the ledges and setbacks of a group
music
of houses, the widths of the single houses and the depths can be determined. The advantage
of the midpoint displacement method over music is that the fractal dimension can be regulated
and determined in the first case.
5.8.2 IFS in Architecture
Another possible application of fractal geometry to architecture is through the

phenomenon of IFS, the iterated function system. Here the rules are more important than
IFS
knowing what will be produced from the very beginning, which means that the results are
influenced by the instructions and not by the primary products. Thus the planning method is
the mechanism, rule or instruction and not the resulting form – which we do not know when
we start the procedure.
In a lecture held by Wolff Plotegg at the Technical University of Vienna he demonstrated

some examples of his own, produced by rules without knowing what the result would be. E.g.
other functions
designing a door normally means that its main function is opening and closing a room.
Everyone has a certain idea what a door may look like, that is because of our experience. The
opening-mechanism is in most cases vertically, the door itself is a thin rectangular object, and
its only function is that of opening and closing. But then Wolff Plotegg thought about other
rules: first the opening mechanism should be horizontal on the ground, the door itself should
be a thick object and he defined another, second function of the door, it should also be a
staircase. The result is not a door the way we all know, but a newly designed one.
5.8.3 Mathematical Fractals

true fractals
As mentioned in chapter ‘5.5.7.b Greg Lynn’, Greg Lynn used a process rule for the basic
shape which led to a mathematical fractal. Only then he translated and adopted it to the
surroundings and the function.

6 Fine Arts and City Planning ‘Fractals and Fractal Architecture’
6 Fine Arts and City Planning
introduction
Whether fine arts and city planning offer fractal geometry or not can once more be proved
by finding characteristics of fractals in the form of paintings and cities, but also by calculating
the fractal dimension with the aid of the box counting method or the scaling range analysis.
After the section about fine arts and an analytical part about cities, a section of some possible
applications of fractal geometry in city planning as a helping tool will be mentioned.
6.1 Fine Arts

Artists have always used geometry in the one or other way for their works, although they
dimensions
may not have been conscious of that. First the decision has to be made if something should
be formed in three or two dimensions. Second there are many rules for composition and
proportion found in design and art that are based on dimensions. Thinking of one of the
most important proportion-rules, namely the ‘golden section’, this turns out to be, as
mentioned before, a fractal sequence. Beside that art can be interpreted as a way for
finding the basics of beauty and harmony that are found in the laws of nature. In this way
chaos and fractal geometry may help to explain and prove the ‘rules’ of beauty.
6.1.1 Description of Two Paintings
comparison
The basic difference between Euclidean and fractal geometry can be shown at two
paintings of the last century which are compared in a purely formal way in the following
section. That is ‘Composition with Black, Red, Grey, Yellow and Blue’ by Piet Mondrian of the
year 1921 and ‘number 8’ painted by Jackson Pollock in 1949.
6.1.1.a Piet Mondrian
‘Composition with Black, Red, Grey, Yellow and Blue’ of picture 62 represents a painting
that belongs to a style of abstract-geometric forms which is very close to Euclidean geometry, Euclidean geometry
like the modern architecture of the 1920ies which asked for Euclidean shapes. In this sense
Mondrian’s painting can be described by straight lines and simple, smooth rectangles. E.g.
one may start on the right upper side with the red rectangle of a certain proportion and relate
all parts of the painting to it, defining width, height and color of each single object.
The painting is located in the development of the style and message of Piet Mondrian.
Out of the organizing principle of the synthetic cubism he developed the arrangement of
areas with certain colors, at which first the motive of the tree turned into an abstract set of
lines. Later on, his paintings lost any reference to real objects.
Nevertheless the process of painting is not accompanied by a straight mathematical

no feelings
scheme, but instead of that Mondrian worked intuitively in front of the canvas. The aim was
to show pure reality through pure arrangement. There should not be any influence of
subjective feelings and ideas. Mondrian tried to achieve universal harmony and
surmounting of individualism, which meant excluding sensuous perception and visible
reality by a reduction to straight lines, rectangles and a few colors.[01]

Later on Piet Mondrian wanted to express through his paintings the complex metaphor for
the encounters, the misadventures and the mutual accomplishments of which life is made
up.[02] By that he turned away from the abstract Euclidean geometry to an abstract ‘Fractal
Rhythm’, which displays the cascade of interest that can be also found in nature. An example
for this rhythm is offered by the painting ‘Victory Boogie Woogie’ from 1944.
color sequence
For describing the series of colors of one strip of the painting – and by that the Fractal
Rhythm –, Carl Bovill simplified the color use to red, blue, yellow, black and , which lead to the
following sequence of colors for this certain strip:
1, 5, 1, 3, 5, 3, 4, 2, 3, 3, 4, 2, 3, 4, 3, 1, 3, 4, 3, 4
with 1=red, 2=blue, 3=yellow, 4=black, 5=grey.
The sequence can then be translated into a graph, in which the resulting plot is similar to
noise and offers a mixture of order and surprise which can finally be measured and expressed
through the fractal dimension, see picture 63.[03]
6.1.1.b Jackson Pollock
‘Number 8’ by Jackson Pollock is one of his abstract compositions in which the color is applied
different kind of painting

to the canvas by a spontaneous movement. The results of these action paintings are harder to
describe in Euclidean vocabulary than the abstract-geometric formed ones by Piet Mondrian.
For his painting Pollock used the method of ‘drip painting’ which allowed a free form
of abstract painting. To get a well-balanced object he put the canvas on the ground and
worked on it from each side not knowing what he was doing exactly in the moment when the
colors were dropping on the canvas – the painting had its own ‘life’. Though this method was
spontaneous, he nevertheless could make corrections and ultimately it was his decision
which part of the canvas he cut out and put on the frame. Highly composed, constructed
compositions and working with the paintbrush confined him too much. Therefore he did
not use any existing forms and motives any more.
6.1.1.c Conclusion
Is the painting ‘number 8’ really a more ‘fractal’ painting? Maybe this can be proved through
an algorithm which is recursively applied to the canvas. A simpler algorithm, similar to the
fractal geometry
Brown motion, may run like that: put the canvas on the ground, step on the canvas and walk
around dropping color on it; the way one takes and the kind of dropping is determined at
random telling whether one moves to the left, to the right or even rotates, using much color or
less, circular, straight or drop-shaped. Of course this does not exactly produce the same
painting but the impression may be similar – we know that there are drops, curved lines, thick
and thin parts on the canvas, see picture 64.[04] Anyway what this painting offers in contrast
to Mondrian’s ‘Composition with Black, Red, Grey, Yellow and Blue’ is a mixture of order,
through circular and straight lines, and surprise, where no pattern is recognized.
generally
Most paintings are situated between those extreme cases. By that paintings which mirror
nature in the one or other way, display a cascade of fractal detail in the same way nature does
– e.g. by the rhythm of the spacing of trees in a wood.

6.2 Cities
‘naturally’ grown and ‘planned’

Most cities look like complex structures to us that can hardly be grasped, both with regard
to form – city-boundaries, traffic-networks – and their distribution or connection of functions –
e.g. living, working, traffic, recreation and administration. This kind of irregular looking cities
are often called ‘naturally’ or ‘organically’ grown, indicating that the geometry of such a city
does not seem to be planned in the large – the form may rather be produced by many detailed
and individual decisions which are coordinated in the small.[06] To tame the ‘chaotic’,
disordered looking structure of cities, planners attempted to bring simple, smooth, visual –
Euclidean – order into the city[07]. This leads to the group of so-called ‘planned’ cities,
involving fast growth with planning on a large scale by one or few authorities.
Most cities are formed by a mixture of both groups, depending on the changing assumptions
most cities
throughout the history of a city: e.g. a newly founded Roman military camp – ‘planned’ –, may
have been changed slowly through the Middle Ages – ‘naturally’ grown –, ..., before the
population explosion of the 19th century once more had caused fast growth – ‘planned’.
6.2.1 Cities and Fractal Geometry
As shown in the previous chapters, fractal geometry is able to describe complex forms,
finding out their underlying order and regularity – self-similarity, simple algorithms –, by
reproducing the real world and not by an abstraction into pure mathematics – ‘clouds are not
applications
spheres’. Therefore fractal geometry offers a good field for application on cities, moreover,
even most of the ‘planned’ cities, using the geometry of Euclid and showing simplicity of form,
have been adapted to their context in more natural ways and therefore also contain some
‘organic’ growth and irregularity.[08] Such applications can be the visualization through
computer models based on fractal geometry, measuring patterns of real cities and their
dynamic simulation by examining city-boundaries, networks, hierarchies, urban texture and
the density of population – if cities belong to fractal geometry then elements of these systems
will be found on different scales: self-similarity.
6.2.2 Brief History examples
Reasons for the first appearance of highly ordered geometric forms in cities were the
previously mentioned rapid physical developments – e.g. worker camps in ancient Egypt,
Greek colonial cities, Roman military camps[09] – and the demonstration of political, religious
or economic power – e.g. Agora, temple complex, palace – see picture 65.[10]
Later in the Middle Ages the concept of grid-based city plans – the ‘geometric principle’ –
was still used if fast development was necessary, such as for military purposes, but also for
trade- and colonial-towns. Apart from this ‘planned’ group, the forms of many cities of the
Middle Ages
Middle Ages were compact but nevertheless irregular, in them buildings were situated around
the central market square and the church, following the form of the terrain – the ‘geomorphic
principle’. The manifold townscapes of that time period were mainly caused by the competition
between different social groups, such as clergy, aristocracy, citizens and craftsmen, who lived
in separate buildings, streets or even districts – ‘many different decisions on a small scale’.
This also led to different expressions of their representation-buildings, such as churches,
chapels, city halls and corporation-buildings.[11]

The end of the flourishing time of Middle-Age-type cities came with the strengthening of
the thoughts of territory and by that with the process of re-feudalism, which rang in a new
epoch – Renaissance and Baroque. Many cities had to defend their freedom and prosperity
against this new power of absolute principality. One of the expressions of this new power
can be found in the idealized city-plans – most of which were orientating themselves on the
10 books of Vitruv – see picture 66.[12]
Renaissance
Beside the influence of the rediscovery of the Greek and Roman architecture, both, the
developments in the architecture of fortification – compact space and fast maneuverability –
and the discovery of perspective – focusing the description on the eye of the viewer, which can
be interpreted by relating the world to man – led to radial focus in city plans and by that
assisted the development of ideal town plans. But only a few of these theoretical plans were
finally realized, because the phase of increase in population and the colonization of the
European continent were mostly finished in the middle of the 14th century – so there was no
need for new towns or an increase of existing ones. Beside that the governments of that time
lacked political stability and the money necessary for such long-time projects.[13]
Then during the period of Baroque the population increased, so that new towns were
necessary to relieve the old ones, whose city-wall-enlargements could not catch all
inhabitants any more. Because of their expression of absolute power, the sovereigns
supported the realization of idealized city-plans. In consequence it was the time of Baroque
in which the views of Renaissance were put into practice: the city as a functional,
Baroque
calculated, homogenous, comprehensive work of art that represents an object of
prestige.[14] Idealized plans were also applied to existing cities by the construction of
idealized fortifications, but also by cutting up the existing structure for installing radial
streets and straight axes. The concept behind idealized city-plans – perspective,
fortification, visualizing power – can today be interpreted, according to the phenomenon of
fractals, as those ‘simple’ underlying rules which produce complex forms – the whole and
the parts following the same principles and thus forming a unity.
The function of self-defense and protection against the surroundings of the idealized city-
plans of the Renaissance and Baroque could be improved by increasing the available – wall self-defense
– space, replacing the straight running fortifications by regular triangular segments, which
allowed shooting at and protecting every segment of the border. The increasing length –
space filling phenomenon – and the roughness of the border, showing indentions and
convexities, indicate formal similarities to the Koch Island. Such similarities are also found
in the regularity of the symmetrical-polygonal plans of Vitruv – see picture 67.
After the period of ‘circular’ idealized cities, planners once more turned back to grid-based
plans. E.g. from the late 18th century onwards the development of American – ‘new’ world –
cities has been dominated by rectangular patterns, though physical constraints and individual
modern ages
decisions have often changed such pure gridirons. But also in the ‘old’ world of the early 19th
century the fast growth of ‘western world’ cities led to a higher extension of planning – on
bigger scale – and by that in matter of speed and convenience to grid-based plans – also
symbolizing a break with the past.[15] Then the geometry of the ideal town has been changed
during the 20th century: it has become more curve linear – keeping still linear nevertheless –,
organized around communication routes – large road systems – with more focus on land uses
than on specific buildings.

6.2.2.a Modern Grid-Planned Cities
In the article ‘Chandigarh, the capital of Punjab’ Helmut Borcherdt describes the ‘planned’
grid planned
city Chandigarh by Le Corbusier as an example of a car-influenced garden-city-grid, where the
building-blocks are situated on like soldiers on the drill ground. Everything is ordered and
separated: different living-quarters for different social classes and separate sections for hotels,
banks and shops – see picture 68. ‘Something beautiful evokes emotions, order by its own
leaves one cold’.[16] This is a cry for a mixture of order and surprise whose expression comes
through fractal geometry and its degree is given by fractal dimension.[17]
Another example of a newer grid-based city is Islamabad by Doxiadis – the new capital
some more examples

of Pakistan.[18] In this case the site-plan of the city is crossed by water. But instead of
reacting to the environment – which would have meant to break the straight lines of the grid
and modify them, leading to a more fractal, ‘naturally’ grown city-model – a lot of expensive
bridges were built to continue the grid across the water, see picture 69. One more example
of rectangular plans that show the lack of harmony between the city and its environment
represents San Francisco, which got its typical character of steep streets by the straight,
continuing grid put on a mountainous site.
What are the differences between these newer examples of ‘planned’ cities and older
‘naturally’ grown ones? The experience of space in older cities was reserved to the resting
and walking viewer: in such cities every step changes the interactions of buildings,
standing close side by side; the pedestrian is able to perceive the spatial structure of the
townscape, the sequences of streetscapes and squares – the main object of perception in
space and time

older cities was space. Newer cities are observed as space-time sequences at high speed
– the spatial structure of the townscape is hardly experienced anymore by man, driving
through the city, see picture 70.[19] It seems that today’s city planning proceeds from the
fastest possibility for moving from point ‘A’ to point ‘B’ in large streets by the fastest vehicle.
In Chandigarh the negative fact of modern-planning is even stronger than anywhere else,
because proportions are not appropriate for human beings and natural behavior has been
excluded: e.g. the square in front of the government-quarter is too large to stay there, it is
too hot to cross it by foot, and cars are parked only in regions of shadow, which leads to
an empty, endless-looking concrete surrounding.[20]
6.2.2.b Natural Growth – Two Hostels of the Yale Colleges by Eero Saarinen
There also exist examples of ‘modern’ ways of planning that produce the character of
Italian mountain village
growing, as illustrated by two new student’s hostels of the Yale Colleges by Eero Saarinen of
the year 1958 – the Stiles and Morse Colleges.[21] In this case, a small footpath – that is lying
between the two new hostels – leads to the tower of an existing strongly symmetric new-gothic
building. Opposite this straight building, Eero Saarinen put a semicircle to define the free-space
as a circus that is enclosed by houses. During the planning he pinned up some of the most
popular squares of Italy on the walls, including the one of Siena, which may be the reason why
the resulting oval place reminds us of this famous one. Beside that, with lots of irregular edges
and angles and different heights the row of buildings takes up the structure of an ‘naturally’
grown Italian city: in contrast to the new-gothic buildings of the college, the student’s hostels got
romantic ground plans that should ‘conjure up’ the feelings of an Italian mountain-village like
San Giminiano – see picture 71.

Eero Saarinen pointed out that parts of a building should follow the strong, simple concept
of the whole – each part has to be an active part of a certain overall theme, which is valid for
‘romantic’ concept
the ground plan, construction-systems, colors inside the house and even door-handles.
Following from that he noticed that modern buildings would not have fit into the historical
environment of Yale-University. Thus Eero Saarinen adapted them to the environment by
using the concept of an old romantic city planning – in consequence everything had to follow
this leitmotif. He called the concept ‘romantic’ in the sense that the emotional value of the
building is higher than the logical morphology – the modern style was refused because it offers
nothing personal and subjective.[22]
simple growing modells

6.2.3 Naturally Grown Cities
The simplified growth model of the ‘organic’ city spreads out from one center of initial
growth or seed in form of waves of development. This first approach can be modified by radial
lines of transportation along which growth often proceeds faster – resulting in star-like shapes
–, but also by the shape of terrain – rivers, hills and the like – and possibly existing defensive
walls – restriction of growth, see picture 72.[23]
For developing more realistic models, patterns of real cities have to be taken into
consideration with respect to their growing through time. But if such a development of cities –
visual history – is analyzed with regard to some deeper order, several problems will arise: e.g.
real patterns
the different quality, detail richness, scale and exactness of the presentation of city-plans,
which have been changed during the centuries and the absence of unbroken availability of
maps make statements more difficult. Beside that, most cities before modern age were small
and compact with higher density caused by slower transport and range. This changed with
better and faster transport-systems and building technologies – since the 17th century the city
walls have gone and the industrial era began –, which led to bigger cities with lower density.[24]
Another aspect that should be taken into consideration is that ‘irregular’ – ‘naturally’ grown
– and ‘regular’ – ‘planned’ – forms vary within the same city with respect to scale – at one scale
scale
the city may belong to the first, on another scale to the second group. That means, if the shape
of the city can be defined through its fractal dimension, then this dimension will change
through scale – see picture 73.
6.2.3.a Form, Structure and Hierarchies
‘Form is the shape of contents’.[25] From this follows that with respect to cities, first – in the
sense that the word ‘form’ refers to the shape, the outward appearance – form represents the
shape and forces
spatial patterns of the physical elements that cities consist of. These elements belong to the
following three categories: ‘networks’ – e.g. transportation networks for people, goods, water,
energy and information –, ‘buildings’ – including residential, commercial and industrial
buildings –, and ‘open space’ – such as parks, gardens, places and courtyards. But, secondly,
the word ‘form’ also involves the forces that create this visual shape, and with regard to that it
is important with the study of form to think about the processes which give rise to it. The first
association of process with form is that of ‘growth’ – the dynamic viewpoint –, involving the
changing of objects through the interactions of forces – ‘organic form’. The second association
is ‘function’ in the sense that processes, containing the forces that produce form, have specific
functions – hence form is understood as the product of functions.[26]

Hierarchies: As following from above, the external form – the shape – can be described by
its internal, invisible form – the structure. The structure is itself composed of elements, so-called
external – internal form

basic components, and their relations, the interactions and functions of elements. With respect
to cities the elements are called units of development – the housing or a ‘block’ –, which are
linked to each other through various communication networks. The related functions of the
elements belong to employment, commerce, education and recreation. All these elements and
relations – structure – of cities cause their complex geometry – shape.[27] In general,
decomposing the system structure, producing subsystems, involving sets of components which
can be arranged according to a hierarchy, in which the subsystems and the elements
respectively, may reflect the same form at different levels: self-similarity – see picture 74.[28]
Cities are organized hierarchically into neighborhoods whose rank and the spatial extent
depend upon the economic function which they offer to the surrounding population –
centers – self-similarity and hierarchy

specialized centers serve larger areas, whereas those of local needs serve smaller ones. This
hierarchy in cities is formed by centers and their hinterlands, which have several elements –
functions – in common, starting with the Central Business District – CBD –, followed by some
district centers, a larger number of neighborhood centers and even more local centers.[29]
Such a hierarchy also exists for the transport system – reaching from primary or trunk down
to local distributors, from freeways to pathways – the educational and the leisure system.
Just the other way round, the self-similarity on different scales, such as represented by the
centers within cities, can be generated if one knows the right functions and rules respectively.
These rules – involving office, government, trade and commercial functions, the range and
multiplicity of goods and services and the quantity of population served – are then applied
recursively on every scale, starting with the largest center, the CBD, leading to different ranks
in the hierarchy of centers.[30]
Classification: The idea of self-similarity and hierarchy can also be carried on by

comparing the classification of rooms in a dwelling with the classification of parts of the city.
On such different scales, there are always sections for leisure, traffic, working and living: e.g.
from the freeway over a footpath to the corridor of the dwelling, from the entrance of the city
to the entrance of a building, from the office-tower to the desk, from the big supermarket over
a smaller food store to the refrigerator.
6.2.3.b Growth Models
The DLA model, introduced in chapter ‘3.4 DLA Models – Diffusion-Limited Aggregation
Model’, can be treated as a baseline model for simulating the growth of cities. The spatial
patterns of cities evolve, similar to the DLA model, by adding cells or basic units – individuals,
households or transportation links, represented by their occupied space – around some
simulation
central point – the CBD. This leads to tree-like or dendritic forms, which show fractal
characteristics such as self-similarity – see picture 75. Batty and Longley mention in their
book ‘Fractal Cities’ that the growth process itself contains codes which determine how the
organization of the basic units might display the self-similarity of forms and functions. That
means if the growth of cities is planned at any scale, the individuals or agencies involved
almost subconsciously take account of economies of agglomeration, the requirement that the
surrounding population is served efficiently by similar functions and services of different order
on different scales, linked by transportation systems – this minimizing costs and effort.[31]

When looking more closely at the growth process of the DLA model and that of cities
some differences appear. First, the structures produced by the DLA model grow irreversibly
in contrast to urban development, where the individuals are free to move, hence it is
possible that already occupied cells become again unoccupied; secondly, the patterns of
problems of comparison
cities are more compact; thirdly the random walk of the cells of the DLA model – before
they are fixed – has no analogy in cities. Beside that, to get a comparable dynamic
structure of real cities in respect to the DLA model, the position of each individual and its
related space respectively has to be known at every time. In the absence of proper data
the structure of real cities can be simplified by dividing the city into a grid – with a lateral
length of e.g. 50X50m –, getting the distribution of the population by census. The
‘occupied’ fields of the grid are then treated like the ‘living’ cells of the lattice of the DLA
model. This of course leads to a more general and coarse image of the city and offers less
dendritic forms. Anyway, with the aid of distance and the density or the number of occupied
cells respectively, the fractal dimension of the city can be measured.
The Dielectric Breakdown Model – DBM – offers a more realistic growth model, dealing
with the relation between fractal dimension and compactness. Compactness can be
DBM model
changed through a certain parameter – lowering the probability of occupying a cell can
increase the compactness.[32] Moreover natural restrictions can be brought into diffusion
by the factor of unlikely or forbidden development in certain areas – e.g. in the case of
mountains. Beside that the development of additional models may produce tunnels or
bridges – see picture 76.
6.2.3.c Urban Boundaries
In general, boundaries are dividing and connecting at the same time, marking the transition
between different uses or territories. From this follows, that urban boundaries can be defined
urban boundaries
as the change-over from the urban to the rural area, e.g. indicated through different land-uses
that reach from high density commercial uses at the CBD to low density agricultural uses at
the edge of the city including differences in their built-up density – villages that are connected
to the urban area by some development may also be related to the city. Beside the definition
through land uses, the population density also gives information about where the separation
of the ‘city’ from its environment takes place – from high density or urban to low density or rural
area.[33] It is obvious that the course of the urban boundary ultimately depends upon the
definition of the changeover, that is which kind of land-use is included and what the limiting
value of population density is respectively.
Anyway the resulting lines look irregular, not smooth in form and are closed, marking out
some area, from which follows, that urban boundaries are more than a one-dimensional –
straight – line and less than the area it encloses – it does not fill the entire plane. Thus the
fractal
fractal dimension may be between 1 and 2, the final value depending upon what is measured
– which is the final definition of the course of the boundary – and which method of measuring
is used – e.g. different results through the structured walk method and the cell count method
–, see picture 77.
Their fractal dimension can compare the boundaries of cities with each other, like
Richardson had done with coastlines and borders of countries – see chapter ‘2.2.1.c
Richardson’. But what are the reasons for the different degree of irregularity at the edge, and

resulting from that for the different fractal dimensions of city boundaries?[34] First, their form
and shape is determined by physical constraints presented by the specific geography of the
area, including steep terrain, the sea and lakes, the course of rivers, which themselves have
often been changed by man – straightening leads to lower fractal dimensions –, and generally
areas that can hardly be built-up – including the dependence upon the conditions of the soil.
influences
Secondly they are influenced by the improvement of building and transportation technology,
which are both connected with the history of the city. Thus the development of the tramway
enabled a better accessibility of the area of the city and the introduction of concrete and the
assembling construction in prefabricated units respectively gave way to cheaper and faster
building and by that to more spacious suburban housing – both lowering the fractal dimension
of the edge. Thirdly patterns of land tenure, the size of building plots and the demand for
residential space also represent possible influences.[35]
In any case, cities can be observed in two different ways: first through their history and
secondly through different scales. The first one informs about the period certain achievements
and their effects like the development of new transportation technologies have lasted or in
general about the time and period changes have taken place, such as the extension of docks
in the course of industrialization, which may have straightened the otherwise rough coastline.
scale and time

The second one offers insight over which range of scale various processes act, indicated by
the Richardson plots: the generally slight curved behavior of the log-log graph, which is
derived from the different effects of processes – operating on different scales – on the
irregularity, is sometimes interrupted by straight line sequences. These sequences expect that
a certain process or better to say, a single set of processes is operating within this specific
range. In general Batty and Longley consider that the fractal dimension is decreasing with
scale as there is greater control on smaller scales, and that the fractal dimension on smaller
scales is decreasing over time, because of greater controls over building technology and land
developments – see picture 78.[36]
6.2.4 Helping Tool
6.2.4.a Mathematical Fractals

application
As indicated before, some mathematical fractals can be used as a visual help for planning
streets, footpaths and the like under the view-point of irregularity or in line with the question
about how much of a certain area can be supplied – the higher the fractal dimension the higher
the irregularity and the more of the entire space can be reached. Besides, fractals may also
act as a first approach for defining the distribution of buildings or the size and position of
properties, the fractal dimension of the resulting site-plan saying something about the
irregularity of the project – see picture 79.
6.2.4.b Curdling
Mandelbrot named the process that produces a fractal dust, which is a disconnected set
chance
of points with clustered characteristics, ‘curdling’. Examples for such a fractal dust are the
Cantor set, introduced in chapter ‘3.1.1 Cantor Set’, as well as the cluster of stars and
galaxies in the night sky, with the latter clustering, in contrast to the first, includes some
randomness. How can the ‘curdling’ process be described? First, the starting image, which

may be a square, is divided into a certain number of pieces – the following example uses
a division into nine pieces that produces a grid of three columns by three rows. In the next
stage, chance decides which of these nine squares or boxes remain and which ones are
removed from the grid. This can be done by a random number generator, or if the random
choice is one half, simply with the aid of a coin. The concept of ‘curdling’ is that the mass
of the material of the original number of squares flows into the remaining ones. In the next
stage, the procedure is repeated for the squares that are kept up – hence each of them is
once more divided into nine smaller squares and again the decision whether the new ones
remain on the grid or not is made by chance. This process can be repeated until infinity
with using the same or different probabilities for keeping up the squares at each stage –
see picture 80.[37]
The fractal dimension of a fractal dust, produced in such a way, can be measured with
the aid of the box-counting method, using the formula Db=[log(N(s2))-log(N(s1))]/[log(1/s2)-
log(1/s1)] – introduced in chapter ‘4.2.3 Box-Counting Dimension ‘Db’’. There ‘Nsn’
represents the number of boxes that remains at a certain grid size ‘sn’ – stage –, which is
defined as the factor of reduction or the reciprocal value of the number of columns of the
fractal dimension
specific grid. Relating the probability to the fractal dimension it turns out that the fractal
dimension decreases when the probability becomes smaller and that there exist some upper
and lower limits. Thus the upper limit is given when all squares remain at each stage, which
means a probability equal to one and a fractal dimension of two. The lower limit of the
example with a division into nine pieces at each stage is given by one ninth, because then
only in the ideal case one square remains at each step until infinity, while in reality no or two
squares may also be likely – if only one square is counted at each step, the fractal
dimension decreases to zero: log1-log1=0. Finally in the case of a probability lower than one
ninth, it is unlikely that any square will remain at some stage.[38]
The ‘curdling’ process, as described in this section, produces a fractal dust that may act
as a first sketch of a site plan for one-family houses or row houses, the fractal dimension
indicating the density. In this connection the environment, a mountain ridge or the like, can
be used as an instruction for the ‘curdling’ process in so far that the probability can be
derived from the measured fractal dimension of this surrounding. But repeating the ‘curdling’
site plan
process with the same starting options, that is probabilities, results in different shapes,
which is also pointed out by different actual fractal dimensions – the theoretical fractal
dimension is generated with the remaining boxes of the ideal case of a certain probability.
Thus a couple of ‘site-plans’ of the same probability can be generated, where finally the
most useful ‘site-plan’ elaborated further by adapting it to the surroundings, such as existing
roads, hills or rivers.

7 Problems with Measuring ‘Fractals and Fractal Architecture’
7 Problems with Measuring

The box-counting method is the most frequently used method for calculating the fractal
problems
dimension of images and plans throughout this paper – see chapter ‘4.2.3 Box-Counting
Dimension ‘Db’’. Hence it is necessary to sum up the various problems that may arise by
using it – e.g. caused by the dependence of fractal dimension upon the quality and
preparation of the image, the sizes of the biggest and smallest boxes used – that is the
definition of the upper and lower scale for measurement – and the starting position of the
boxes in relation to the image.
7.1 The Size and the Quality of an Image

First of all it is necessary to remember that the image or the plan of which the fractal
the plan and the distance

dimension is finally calculated only represents a simplification of the real object, e.g. of the
front of a building. In this connection plans of smaller scales generally display more details
of the building than those of bigger ones. This is quite similar to the distance of the
observer to the real object: if one stands far away, something can be hardly recognized,
except one uses a pair of glasses – compare picture 28. Getting closer, which means
increasing the scale of the plan, the details become visible – certain details can only be
identified from a certain distance or scale on. Thus translating an object into a plan first of
all means to define the distance- or the scale-range – with the aid of the biggest and
smallest recognizable detail. Using box-sizes beyond this range, in connection with the
measurement of the fractal dimension by the box-counting method, this leads to falsified
results – no additional details of the real object, the building, are included into the
calculation because the plan does not contain them.
Looking at the Koch curve and other mathematical fractals this would mean that,
the distance and the iteration

depending upon the distance they are observed from, the number of iterations can be
defined. E.g. from far away maybe only the first three iterations will be distinguishable,
and all iterations from then on do not seem to change the image any more. Up to a certain
scale the calculated box-counting dimension will give some conclusion on the fractal
dimension of the real object. But then, if too small boxes are used for measurement, that
is the scale-range of observation remains too low, the image will quickly turn out to be a
one-dimensional line. From then on no additional details can be picked up because the
image does not contain them – the length of this curve is limited in contrast to the real
Koch curve.
Some other influences represent the quality of the image itself, which may be
some more influences
translated into the language of the computer by the number of pixels per inch – high or
low definition. ‘Quality’ also means how many details are getting lost only because of the
preparation of the plan – by scanning and after-treatment – and not because of the
distance, that is scale. Then the interpretation of the wall of floor plans, whether they are
painted black, white or grey, influences the results of the measurement. But also the
contents of the plan changes the fractal dimension: elevations may also display the
environment – natural or built one –, floor plans may contain tiles on terraces and/or
sanitary fixtures and other equipment – some closer look at this theme can be found at
picture 81. Such differentiations are taken into consideration when analyzing house types

in chapter ‘99 Statistics’. In general images that represent elevations and floor plans of
buildings have to be similar in the way they pick up details, in their quality and in their
scale if they should be compared with each other.
7.2 The Range of Box-Size

The absolute smallest possible box-size that can be used for calculating programs
represents one pixel, as a resolution limit of the computer and of the digital image respectively.
But in general too small box-sizes would mean that every difference caused by the preparation
lower limit
for the computer is also taken into consideration. In consequence the thickness of a line of the
image should represent the absolute lowest limit size for the box-counting method. That means
if the scale range we are looking at is between one and two pixels and the line is three pixels
wide, the dimension of this line on the lower scale increases approaching two, though a line is
one dimensional in Euclidean sense. This and the fact that below this scale no more details can
be picked up, are the reason why the lowest box-size should be bigger than the thickness of
the lines – see pictures 82-84.
The other extreme, the biggest box-size, can be as large as the image itself. But this extreme
can be eliminated, because when counting the number of occupied boxes at this stage, it does
upper limit
not make any difference whether there is only one black pixel or many, the result is always one.
Only some stages later, decreasing the box sizes, the difference between one pixel and many
will be recognized by a low fractal dimension for the first and a higher one for the latter case.
So often the first, biggest, box-sizes are excluded from the measurement.
The box-sizes in between the two extremes, the smallest and largest box-size, should
factor two
generally decrease by the factor two from scale to scale. So the outer borders of the
measurement-grid remains the same, and by that the relative position from one box-size to
the previous one – see pictures 82-84.
7.3 Starting Points

There is more than one possible starting-position for the boxes to be laid over the image.
E.g. the left bottom corner of the box may be identical with the left bottom corner of the image.
But it is also possible to start three pixels left and two pixels down of the corner of the image.
left lower position
In this connection pixels are used because most measurements of this paper were made with
the computer and so the lower limit of resolution is one pixel. Thus if the box-size is only one
pixel then the starting point has only one possible position on the bottom left side and each
box of this grid is exactly one dot of the image. Whereas if the box-size is 2 by 2 pixels, four
possible random origins are possible: first all four pixels are situated on the image – the left
lower pixel of the box is identical with the left lower pixel of the image –, secondly only one
pixel is located on the image – the origin of the box is one pixel below and one left of the image
–, thirdly the lower half of the box is outside the image and fourthly the left half sticks out. If
the box-sizes increase, there are of course even more possibilities for the position of the
starting point of the box – see picture 83.
How do computer-programs manage such problems? E.g. the computer program ‘Benoit’
starts with a defined side-length of the largest box. The user can change this starting size,

but usually this is given by one-fourth of the width or height of the image. Beside that he
can also define the decreasing coefficient of the box-size. To minimize the error deriving
from the box-sizes and starting positions, the program defines a certain increment of grid
‘Benoit’
rotation - the grid rotates with a certain angle. For each grid-position – rotation – the
occupied boxes are counted. The computer-program then calculates the minimum number
of boxes occupied by the image for each box-size – following from that it does not matter
whether the boxes exceed the image or not.[01]
7.4 The Slope of the log-log Graph

The slope of the log(N(s))-log(1/s) graph defines the fractal dimension of the object, which
the trend of the slope

is getting more exact, the more results – N(s) versus 1/s – are available. Points of the graph
that are situated beside this slope – line – may indicate that first the image of observation is
not a fractal, secondly the box-sizes are defined as too big or too small, thirdly the image-
quality is bad or fourthly as mentioned in the previous chapter the image is multifractal. In the
analysis of house-types of chapter ‘9 9 Statistics’ this circumstance is brought in by the position
of the points of the graph in relation to the slope: the results are called ‘very smooth’ – the
points are congruent with the slope –, ‘smooth’ – indicating minimal deviation –, ‘diverging’ or
‘strongly diverging’. The last two cases are excluded from measurement because it is not clear
whether elevations are multifractal or if the deviation results from the influences as described
before, like bad quality, wrong preparation and so on.
7.5 Programs
different programs
The measurement of the fractal dimension of one and the same image with different
computer-programs may lead to different results – if the same image with the same quality
and size is run through different programs, this will offer the differences in computing-
adjustments. Some of the differences of the programs are given in picture 84 where I have
used different computer-programs – ‘Fractal Dimension Calculator’ and ‘Benoit’ – for the
same calculations. In comparison and for better understanding I have also measured the
box-counting dimensions of some images by hand. ‘Fractal Dimension Calculator’
The instruction of the ‘Fractal Dimension Calculator’ by Paul Bourke compares some
images of known fractals with the results of the dimensions calculated by his program.
Thus the Koch Island has a measured fractal dimension of 1.5, while the box-counting
dimension is 1.53 – the difference between these two results is 1.8%. The Koch ‘Coastline’
has a measured fractal dimension of 1.262 and a box-counting dimension of 1.22, the
difference being 3.6%. In the first case of the Koch Island the slope and the curve of the
points of the log-log graph are quite congruent, while the curve of the Koch ‘Coastline’
differs a little bit from the straight unit-line. For a Euclidean line the measured dimension is
0.994 instead of 1, so there is a deviation of only 0.6%.[02]

10 Appendix ‘Fractals and Fractal Architecture’
Literature
The following section gives a choice of literature for the main as well as for the analytical part.
For main part:
Books
Arnheim Rudolf, Kunst und Sehen: e. Psychologie d. schöpfer. Auges (expanded and revised edition 1978), Walter de
Gruyter & Co., ISBN 3-11-006682-3
Baedecker, Frankreich (1988) 4. Auflage, Karl Baedecker GmbH Deutschland
Batty and Longley, Fractal Cities (1994), Academic Press Inc., ISBN 0-12-4555-70-5
Bax Marty, Complete Mondrian (2001), V+K Publishing, Blaricum
Benevolo Leonardo, Die Geschichte der Stadt (8. edition 2000), Campus verlag GmbH Frankfurt/New York, ISBN 3-
593-36439-5
Briggs John and Peat F. David, Die Entdeckung des Chaos (6.edition 1999), Deutscher Taschenbuchverlag, ISBN 3-
423-33047-3
Brockhaus, Der große Brockhaus Kompaktausgabe in 26 Bd., Band 1-26 (1983) aktualisierte 18. Auflage, F.A.
Brockhaus Wiesbaden
Bruckgraber Iris and more, Kunst des 20. Jahrhunderts - Museum Ludwig Köln (1996) Originalausgabe, Taschen
Verlag GmbH Köln, ISBN 3-8228-8819-2
Borcherdt Helmut, Architekten – Begegnungen 1956-1986 (1988), Georg Müller Verlag GmbH, ISBN 3-78844-2181-4
Bovill Carl, Fractal Geometry in Architecture and Design (1996), Birkhäuser Bosten, ISBN 3-7643-3795-8
Ching Francis D.K., Bilderlexikon der Architektur (1996), Campus Verlag, ISBN 3-593-35579-5
Freytag & Berndt, Unterstufenatlas (1979), ISBN 3-85084-050-6
Fuhrmann Peter, Bauplanung und Bauentwurf: Grundlagen und Methoden der Gebäudelehre (1998), W. Kohlhammer
GmbH Stuttgart
Gössel Peter and Leuthäuser Gabriel (editors), Frank Lloyd Wright (1994), Benedikt Taschen Verlag GmbH, ISBN 3-
8228-9754-X
Jencks Charles, Die Architektur des springenden Universums, Arch+ number 141 (1998), Arch+ Verlag GmbH
Aachen, ISSN 0587-3452
Jencks Charles, The Architecture of the Jumping Universe, (1995), academy editions, ISBN 1 85490 406 X
Joedicke Jürgen, Architekturgeschichte d. 20. Jahrhunderts (1990), Karl Krämer Verlag, ISBN 3-7828-0459-7
Jürgens Hartmut, Peitgen Heinz-Otto, Saupe Dietmar: Fraktale – eine neue Sprache für komplexe Strukturen,
Spektrum der Wissenschaft (9/1989)
Jürgens Tietz: Geschichte der Architektur des 20. Jahrhunderts (1998), Könemann, ISBN 3-8290-0512-1
Koch Wilfried, Baustilkunde – Das große Standardwerk zur europäischen Baukunst von der Antike bis zur Gegenwart,
Mosaik Verlag
Koepf Hans, Baukunst in 5 Jahrtausenden (10. edition 1990), Kohlhammer, ISBN 3-17-011072-1
Lynn Greg, Das erneuerte Neue der Symmetrie, Arch+ number 128 (1995), Arch+ Verlag GmbH Aachen, ISSN 0587-
3452
Mandelbrot Benoit B., Dr. Zähle Ulrich (editor of the German edition), Die fraktale Geometrie der Natur (1991)
einmalige Sonderausgabe, Birkhäuser Verlag Berlin, ISBN 3-7643-2646-8
Müller Werner and Vogel Gunther, dtv-Atlas Baukunst, Bd.1 (12. edition 2000), Deutscher Taschenbuch Verlag GmbH
&Co.KG München, ISBN 3-423-03020-8
Müller Werner and Vogel Gunther, dtv-Atlas Baukunst, Bd.2 (11. edition 2000), Deutscher Taschenbuch Verlag GmbH
&Co.KG München, ISBN 3-423-03021-6
Peak David and Frame Michael, Komplexität – das gezähmte Chaos (1995), Birkhäuser Berlin, ISBN 3-7643-5132-2
Pevsner, Honour, Fleming, Lexikon der Architektur (1992), ISBN 3-7913-2095-5
Pothorn Herbert, Das große Buch der Baustile (1998), Cormoran, ISBN 3-517-07909-X
Puppi Lionello, Andrea Palladio – Das Gesamtwerk (1994), Paperbackausgabe, Deutsche Verlags-Anstalt GmbH,
ISBN 3-421-03060-X
Schönfeld Jürgen W., Gebäudelehre (1992) 2. überarbeitete Auflage, Kohlhammer, ISBN 3-17-010558-2

appendix ‘Fractals and Fractal Architecture’
Seibt Ferdinand, Glanz und Elend des Mittelalters (1987), Siedler, ISBN 3-88680-279-5
Storrer William Allin, The Frank Lloyd Wright Companion (1993), The University of Chicago Press, ISBN 0-226-77624-7
Toman Rolf, Die Kunst der Gotik (1998),Könemann, ISBN 3-89508-313-5
von Lengerke Christa, Malerei heute – von Pollock bis Warhol, in: Walther Ingo F. (editor), Museum der Malerei,
Edition Atlantis
von Meiss Pierre, Elements of architecture, from form to place (reprinted 1998), E & FN Spon, ISBN 0 419 15940 1
Voß Herbert, Chaos und Fraktale – selbst programmieren (1994), Franzis-Verlag GmbH Österreich, ISBN 3-7723-7003-9
Walther Ingo F. (editor), Malerei heute von Pollok bis Warhol, Sonderausgabe für Atlantis Verlag Zürich, ISBN 3-7611-
0707-2
Zerbst Rainer, Antoni Gaudí (1993), Benedikt Taschen Verlag GmbH, ISBN 3-8228-0067-8
Websites
http://www.asn-linz.ac.at/iaac/gmunden/chaostheorie.htm (04.01.1999)
http://nb.in-berlin.de/scut/essay/frak/mandel.html (07.06.1996)
http://stud4.tuwien.ac.at/~e9426503/soinfoges/mathematik.html (24.09.2001)
For analysis
Benevolo Leonard and Albrecht Benno, Grenzen – Topographie, Geschichte, Architektur (1995), Campus Verlag,
ISBN 3-593-35362-8
Schneider Friederike (editor), Grundrißatlas (1994), Birkhäuser Verlag, ISBN 3-7643-2625-5
Chitham Robert, Die Säulenordnung der Antike (1994), Fourier Verlag Wiesbaden, ISBN 3-925037-77-2
Gellner Edoardo, Alte Bauernhäuser in den Dolomiten (1989), deutschsprachige Ausgabe, Verlag Georg D.W.
Callwey München, ISBN 3-7667-0946-1
Computer programs
http://www.mhri.edu.au/~pdb/fractals/fractdim (15.04.1999); User manual for a mac program.
‘Benoit version 1.2’ TruSoft Int’l, Inc. – Copyright (C) 1997, 1999.
‘Fractal Explorer 1.21’ – Copyright (C) 1998-2000 Sirotinsky Arthur and Olga Fedorenko, Kiev, Ukraine – see picture
15 IFS – ‘A fern and a snowflake’. http://skyscraper.fortunecity.com/binary/34/Index.html (april 2001).
‘Fractal Forge Version 2.3’ by Uberto Barbini – see picture 14 ‘Mandelbrot set and Julia sets’.
‘Lindenmayer-Fraktale 2.1’ – Copyright (C) Ulrich Schwebinghaus 1999-2001 ‘geschrieben in Delphi 4.0’ – a random
sequence was put in the program to get the natural looking coastlines of picture 12.
‘VMbrot’ – see picture 14 ‘Mandelbrot set and Julia sets’ – orbits.
‘Vision of Chaos’ – Freeware (C) Version 30.9, 8th May 2001, created by Jason Rampe – see picture 18 ‘DLA’ –,
http://fractalchaos.freeyellow.com (may 2001).
‘Winfract’ – Fractint for Windows – Copyright (C) 1990-93, The Stone Soup Group, Winfract (Fractint for Windows)
may be freely copied and distributed, but may not be sold – see picture 07 ‘The Mandelbrot set’.
Computer programs for calculating fractal dimension

‘Benoit version 1.2’ TruSoft Int’l, Inc. – Copyright (C) 1997, 1999.
‘Fractal Dimension Calculator’ – Paul Bourke
‘Fractal Dimension – Ver. 1.1’ – bar ilan university
‘Program to Calculate Fractal Dimension – ‘DOS’’ – M. Laubscher

Footnotes
1 Introduction
01 Comparison of natural objects with Euclidean geometry by the mathematician Benoit Mandelbrot in ‘The Fractal
Geometry of Nature’.
Bovill Carl, Fractal Geometry in Architecture and Design (1996), Birkhäuser Bosten, ISBN 3-7643-3795-8, p.4,
Mandelbrot Benoit B., Dr. Zähle Ulrich (editor of the german edition), Die fraktale Geometrie der Natur (1991)
einmalige Sonderausgabe, Birkhäuser Verlag Berlin, ISBN 3-7643-2646-8, p.13.
02 Warning of Michael Barnsley in the introduction to his book ‘Fractals Everywhere’ (1988), Academic Press.
Bovill Carl, Fractal Geometry in Architecture and Design (1996), Birkhäuser Bosten, ISBN 3-7643-3795-8, p.3.
03 The Latin adjective ‘fractus’ – from which the term ‘fractals’ is derived – was used by Benoit Mandelbrot to denote
a phenomenon with a new word. The Latin verb ‘frangere’ means ‘break into irregular fragments’, Mandelbrot
Benoit B., Dr. Zähle Ulrich (editor of the german edition), Die fraktale Geometrie der Natur (1991) einmalige
Sonderausgabe, Birkhäuser Verlag Berlin, ISBN 3-7643-2646-8, p.27.
04 Maybe it is this beautiful interesting appearance that brings a better access to mathematics: the theory of fractals and
chaos is not only an abstract theory but it can also be visualised – mostly through the computer – which surely is
helpful for better understanding of the underlying mathematics.
05 Batty and Longley, Fractal Cities (1994), Academic Press Inc., ISBN 0-12-4555-70-5, p.31.
06 That is what architects may be interested in: forms – their mathematical explanation and their application to
architecture. Today we use the mathematical explanation of complex geometries, fractals, in animated cartoons and
scienc fiction movies to generate trees, ferns, mountains, unknown worlds and whatever is found in nature. The
fractal geometry reduces the complex forms to much simpler mathematical expressions which make computer files
smaller and, reduced to a few iterations, above all quicker to compute – which means to produce. So one field of
application may be in CAD where complex forms of nature such as clouds and trees may be constructed without
using strange complicated wire models or 2-dimensional pictures.
07 Peak David and Frame Michael, Komplexität – das gezähmte Chaos (1995), Birkhäuser Berlin, ISBN 3-7643-
5132-2, p.19
Complexity in architecture means that elements of e.g. a facade – columns, capitals, base, architraves etc – can
be grouped in a way that there is more than one equivalent interpretation presented to the observer. This is true
for Michelangelos’ facade for the funerary chapel of San Lorenzo in Florence (1516-1534). Complexity does not
mean complicated forms but the interconnection of elements that produce the whole. Von Meiss Pierre, Elements
of architecture, from form to place (reprinted 1998), E & FN Spon, ISBN 0 419 15940 1, p.45.
08 Comparison by Benoit Mandelbrot in ‘The Fractal Geometry of Nature’ – Die fraktale Geometrie der Natur (1991)
09 Of course this will not lead to an automatic architect: we know the limits of artificial intelligence. There exists, for
example, a computer program which produces ground plans of villas by Palladio through algorithms. The grammar
of buildings by Palladio was analyzed and brought into rules – for example the repeating symmetric axes and the
continuous main-room. With the set of rules and the repeating recursive application of them, a defined grid and
additional rules for the porticos, windows and doors under the control of Palladios rules of forms, the computer
works out a couple of results that look like the villas built by Palladio – the underlying structure, the character is
reproduced well. But there are some limits shown, for example in the third dimension, where too many rules for
computing would be needed. Beside that the machine has a lack of knowing the meaning of functions – the relation
to the neighborhood is missing. The program is neither able to sort out ‘wrong’ ground floors which are not fitting
that well, nor is there a limit in using the program for more complex forms.
10 The form of the roof of the Casa Batlló by Antoni Gaudí looks like the back of a dinosaur, the columns of the ground
floor seem to be the feet of elephants and the balconies look like faces of animals. Jürgens Tietz: Geschichte der
Architektur des 20. Jahrhunderts (1998), Könemann, ISBN 3-8290-0512-1, p12.
2 Fractals – A Definition
01 Batty and Longley, Fractal Cities (1994), Academic Press Inc., ISBN 0-12-4555-70-5, p.3.
02 Fractals can be called the geometry of chaos. Batty and Longley, Fractal Cities (1994), Academic Press Inc.,
ISBN 0-12-4555-70-5, introduction p.v.
03 Between the range of 20 centimeters and 20 meters – there is not any absolute value for that range – the
influences of man-cultivated borders will be high, which means that the length will not increase very much between

these scales. Above the 20 meters limit the man-formed borders no more have that much influence because then
they nevertheless follow the original natural form and below about 20 centimeters the measurement is again
influenced by the irregularity of stones. This limitation of scale is also used in the analyzes of elevations and ground
plans of buildings. Mandelbrot Benoit B., Dr. Zähle Ulrich (editor of the german edition), Die fraktale Geometrie der
Natur (1991) einmalige Sonderausgabe, Birkhäuser Verlag Berlin, ISBN 3-7643-2646-8, p.39.
04 Spain is a larger country than Portugal, which means it would need more maps of the same measuring scale as
Portugal to show the whole country. Voß Herbert, Chaos und Fraktale – selbst programmieren (1994), Franzis-Verlag
GmbH Österreich, ISBN 3-7723-7003-9, p.13.
05 Mandelbrot Benoit B., Dr. Zähle Ulrich (editor of the german edition), Die fraktale Geometrie der Natur (1991)
Another example for a different border-length is that of the Netherlands and Belgium, where the differences in
length, given in the lexica of these neighboring countries differ by about 20%. This results from a different choice of
unit-length, which means that the smaller country measures its border more carefully on maps of a smaller scale.
einmalige Sonderausgabe, Birkhäuser Verlag Berlin, ISBN 3-7643-2646-8, p.40-41.
08 Bovill Carl, Fractal Geometry in Architecture and Design (1996), Birkhäuser Bosten, ISBN 3-7643-3795-8, p.15
Jürgens Hartmut, Peitgen Heinz-Otto, Saupe Dietmar: Fraktale – eine neue Sprache für komplexe Strukturen,
Spektrum der Wissenschaft (9/1989), p.54.
09 Peak David and Frame Michael, Komplexität – das gezähmte Chaos (1995), Birkhäuser, ISBN 3-7643-5132-2, p.41
11 Though maps of different scales differ in their details, they have equal characteristics: smaller and bigger bays or
inlets are rough outlines and, besides the scale, geometrically identical.
12 Jencks Charles, The Architecture of the Jumping Universe, (1995), academy editions, ISBN 1 85490 406 X, p.44.
13 Gaston Julia like his rival Pierre Fatou analyzed the phenomenon of feedback. They realized the influence of the
constant ‘C’ but they did not have the possibility of computers to generate pictures of its behavior. In simple cases
points near the zero point converge to a certain point – fixed-point of f(z) –, while outer points approach infinity. In
between those two areas there is an infinitely small border that is today called Julia set. Points of the two areas
tend to stay away from this infinitesimal border, to outer or inner areas. Jürgens Hartmut, Peitgen Heinz-Otto,
Saupe Dietmar: Fraktale – eine neue Sprache für komplexe Strukturen, Spektrum der Wissenschaft (9/1989), p.59
Gatson Julia wrote his knowledge down as a war-injured person in a military hospital in 1918. Both, Julia and
Fatou published their knowledge in mathematical sentences which remained forgotten after their publication in
1918 until Benoit Mandelbrot found them again. Their work had been done by their great imaginative power which
is now brought down on the computer screen and thus visible for everyone.
14 Benoit Mandelbrot was born in Warsaw in 1924. In 1936 he emigrated to Paris where he studied at the Ecole
Polytechnique from 1945 to 1947 – 1948 M.S. Aeroscience in Caltech, Pasadena, 1952 Ph.D. in mathematics at
the University of Paris, 1958 moved to USA, member of the research department and IBM-fellow in Yorktown
Heights in 1974, professor of practical mathematics at Harvard University in 1984. Benoit Mandelbrot has been
awarded the F. Barnard-medal of the Columbia-University, a very rare honour, for the development of his fractal
geometry of nature. Mandelbrot from the Thomas-J.-Watson-search department of IBM in Yorktown Heights, New
York, has brought a new way of thinking into mathematics and natural science by his concept of fractals. This
concept was written down in his book ‘The Fractal Geometry of Nature’ published in 1977.
15 There are different kinds of behavior to be seen on the screen:
1. If the brightness of the screen is low and the light in the room is shortly turned on, the thus produced
pattern disappears and the screen remains dark.
2. If the brightness of the screen is increased, sometimes a short flash is produced which turns into a
constant or pulsating pattern.
3. Depending on the adjustments – sharpness, focal length, ... – the pattern is never stabilized and the
light moves around the screen before it disappears, then a new flash dances again around the screen
and disappears ... The sequence is never or sometimes only after a long period repeated.
4. Depending on the adjustments sometimes a spontanous flash produces ‘organic’ structures, moving
patterns with a highly complex spatial structure.
The first and second possibility has too much and the third too little ‘order’. The fourth kind produces interesting
patterns with self-organisation. Peak David, Michael Frame, Komplexität – Das gezähmte Chaos (1995), Birkhäuser
Berlin, ISBN 3-7643-5132-2, p.22-29.
16 Often a regular system turns into a irregular one at once. In 1963 E. Lorenz found that phenomenon when he
looked at the example of mathematical climate models: the wing-stroke of a butterfly in Singapore is magnified
eventually into a hurrican in Florida. The reason lies in the fact that because of the positive feadback very small

power-differences – e.g. the wing-stroke of a butterfly – can lead to enormous consequences. Naturally such
disturbances are compensated by other effects and therefore chaotic systems nevertheless fluctuate between
certain maximum and minimum values. Jencks Charles, The Architecture of the Jumping Universe, (1995),
academy editions, ISBN 1 85490 406 X, p.29.
17 Chaotic systems are deterministic in the way that certain rules influence dynamic behavior. If we know all physical
laws of a part of a chaotic system we can prove these laws in the section – but the complex whole system seems
to be completely random. http://www.asn-linz.ac.at/iaac/gmunden/chaostheorie.htm (04.01.1999).
18 In the past, mathematicians and physicists analyzed and proved laws only of closed sections of a system and drew
conclusions from that with regard to the whole. Repeating the experiment with the same – similar – causes mostly
had the same – similar – effects, the laws were proved.
The condition of a human heart can be described as a deterministic chaotic system. The rhythm of the heart swings
a little bit all the time which can be seen in the curve of an ECG. This medically measured curve itself has fractal
attributes. An ill heart is found because it has a too regular rhythm. The oszillation of the heart ventricles on the
other hand shows total disorder. Peak David and Frame Michael, Komplexität – das gezähmte Chaos (1995), Birkhäuser
Berlin, ISBN 3-7643-5132-2, p.119.
3 Different Fractals
01 Georg Cantor, a German mathematician, created the Cantor Set in 1883. He worked on the foundation of set theory.
02 The Cantor Set once was the solution for a problem that was analyzed by computer engineers like Benoit
Mandelbrot. The analogic signals of the telephonic computer-data transfer were overlapped by extraneous noises
in certain intervals. So parts of the data were useless. Mandelbrot found out that these extraneous noises followed
the scheme of the Cantor Set and so he could solve the problem.
04 Waclaw Sierpinski (1882-1969) was a Polish mathematician – he created the Sierpinski Gasket in the year 1919.
05 The Swedish mathematician Helge von Koch constructed the Koch curve first in 1904, again a long time before
fractals were defined.
06 The Peano curve was constructed by Giuseppe Peano in 1890.
07 Jürgens Hartmut, Peitgen Heinz-Otto, Saupe Dietmar: Fraktale – eine neue Sprache für komplexe Strukturen,
08 http://nb.in-berlin.de/scut/essay/frak/mandel.html (07.06.1996).
09 Programs which compute such a Mandelbrot set assign each pixel of the viewing screen to a complex number. If
each of these points were computed with infinite iterations, it would take infinite time. Therefore the number of
iterations is for example reduced to 500 or 1000 steps. Then it is checked if the absolute number of ‘Zn’ grows to
infinity on high possibility, that is if Zn>2. In this case it can be said with high probability that ‘Zn’ is not an element
of the Mandelbrot set.
10 For finding the Julia set it is also possible to reverse this procedure. This transformation has the effect that an outer
or inner point aims at the Julia set. What does the new transformation then look like? As the cubic root of real but
also of complex numbers can be positive or negative, there are two possible formulas: f1(u)=+√(u-C) and f2(u)=-
√(u-C). These non-linear transformations are then used in the copy-machine, which applies certain rules,
transformations, to a starting image – circle, square, or any other picture. The difference to linear fractals lies in
the circumstance that in general straight lines are not duplicated on straight lines any more but on curves. Jürgens
Hartmut, Peitgen Heinz-Otto, Saupe Dietmar: Fraktale – eine neue Sprache für komplexe Strukturen, Spektrum
der Wissenschaft (9/1989), p.60.
11 Any Julia set belongs to a certain value of ‘C’ which in turn represents a specific point of the Mandelbrot set.
12 IFS which produce forms that look like real plants work like the copy machine, introduced in picture 02. In this
sense they belong to the same category of true mathematical fractals as the Sierpinski Gasket.
14 The equation for the movement of a point along a line is defined by X(n+1)=AxX(n)+B which leads to different results
depending on the chosen values for A and B. E.g. the starting values X(0)=4 and A=-1, B=0 result in X(0)=4, X(1)=-
-1x4+0=-4, X(2)=-1x-4+0=+4, X(3)=-1x4+0=-4, X(4)=--1x-4+0=+4, and so forth. This means that the transformation
jumps between two points: +4 and -4. Changing the strating values to X(0)=+4, A=-0.25, B=1, the results are then
turning to: X(0)=+4, X(1)=-0.25X4+1=0, X(2)=-0.25X0+1=+1, X(3)=-0.25X1+1=+0.75, X(4)=-0.25X0.75+1=+0.8125,
X(5)=+0.7969, X(6)=+0.8008, X(7)=+0.7998, X(8)=+0.8000 and so forth. As this set of results show, this second

transformation is slowly approaching a fixed point around the value of 0.8. This is then called the attractor in the
vocabulary of IFS. Finally the transformation that moves a point on a plane is given by two equations because
there are two coordinates for each point on the plane. The function then looks as follows f(X,
Y)=(X(n+1)=AxX(n)+BxY(n)+E, Y(n+1)=CxX(n)+DxY(n)+F). Bovill Carl, Fractal Geometry in Architecture and Design
(1996), Birkhäuser Bosten, ISBN 3-7643-3795-8, p.54.
15 Bovill Carl, Fractal Geometry in Architecture and Design (1996), Birkhäuser Bosten, ISBN 3-7643-3795-8, p.48.
17 Brown movement: when we look at a glass of fluid which is in balance it seems that all particles are immovable.
Now if we put any tiny object into the fluid, this object will sink to the ground and will stay there without moving. If
this object is a heavy ball it will sink vertically and will not come up again. This is true for objects we can see and
we are used to. When we zoom more closely to the fluid with the help of a microscope and put some little particle
in the fluid we will see another phenomenon: observing it we will see that the particle does not sink down vertically
but moves up and down without any order, rotates without resting. Drawing the way on a paper we can define the
position of the particle every e.g. 30 seconds and connect them by straight lines. In each of these points a tangent
can be drawn, but in reality the particle moves in between, too. That means that the tangents are only true for the
curve of ‘every 30 seconds’ and not for the real course. Beside that the way of the Brown-movement nearly fills
the plane – this would mean that the way of the Brown-movement is topologically a curve of the dimension 1, but
because it nearly fills the plane it is a fractal of higher dimension. Mandelbrot Benoit B., Dr. Zähle Ulrich (editor of
the german edition), Die fraktale Geometrie der Natur (1991) einmalige Sonderausgabe, Birkhäuser Verlag Berlin,
ISBN 3-7643-2646-8, p.24.
18 Voß Herbert, Chaos und Fraktale – selbst programmieren (1994), Franzis-Verlag GmbH Österreich, ISBN 3-7723-
7003-9, p.92.
20 The word ‘attractor’ originates from the Latin ‘attrahere’. Voß Herbert, Chaos und Fraktale – selbst programmieren
(1994), Franzis-Verlag GmbH Österreich, ISBN 3-7723-7003-9, p.26.
21 In the geometrical expression a strange attractor is a fractal.
4 Dimensions
01 Euclid, Greek Eukleides, was a Greek mathematician in the 4th/3rd century B.C. He was the author of ‘Stoicheia’,
‘elements’, which was translated into Latin in late antiquity and is the most important mathematical education
manual. He taught at the Platon Academy of Alexandria.
02 Euclidean geometry knows only a few elements such as line, circle, planes, cubes.
03 A fractal curve has infinte indentations which means that looking at the curve from some distance we may define
a point ‘on’ the curve but with zooming in new details, that is inlets, will come up and will show that we did not catch
the curve where we expected. The point may be on the one inlet or the other which could not be distinguished on
the larger scale.
04 The length of the coastline grows by decreasing the scale of the map which makes the image more exact. Finally
looking at a fractal curve, the total length of the curve cannot be given, because after infinite iterations the total
length of a fractal is infinite.
7003-9, p.315.
06 Fractal dimension, that is to say a non-integer dimension, has been known since 1919, with a mathematical
background by Hausdorff. Felix Hausdorff (1868 - 1942) was a mathematician and author.
Benoit Mandelbrot, in co-operation with biologists, physicists, statisticians, technicians, astrologers and
meteorologists, came to the conclusion that many phenomena which have been described insufficiently up to now
are self-similar as their underlying principle. The geometrical view of them is ‘irregular’ but often also very regular.
In this connection Benoit Mandelbrot called all non-smooth sets fractals or fractal sets, but he really only meant
objects that are in the one or other way self-similar. Such an object does not have any characteristic size because
zooming in means that it turns into itself again over at least a certain scale and for essential characters. The fractal
dimension helps to classify and describe such phenomena.
Fractal dimension provides a quantifiable measure of the mixture of order and surprise in a rhythmic composition.
08 The smallest ‘rn’ for natural structures also has to be choosen carefully because at one stage the scale of ‘rn’
becomes as small as the scale of the image itself and no increase in length would be observed. This is then called
the lower scale of the object.

09 The edges of an elevation may be found around windows, doors, walls, the roof and around certain details, which
depend on the scale of the analyzed plan.
5 Fractals in Architecture
01 Frank Lloyd Wright used the vocabulary of nature not just by forms rebuilt in an organic matter but by translating
the natural structure which mostly belongs to fractal geometry.
03 Mandelbrot Benoit B., Dr. Zähle Ulrich (editor of the German edition), Die fraktale Geometrie der Natur (1991)
05 The elevation informs us about material, in some respect of what happens inside – staircase, kitchen, office –, the
lightness of the rooms – through windows –, maybe about construction, forms, functions, but always in connection
with knowledge of similar types, materials, forms and logical relations.
07 Bovill Carl, Fractal Geometry in Architecture and Design (1996), Birkhäuser Bosten, ISBN 3-7643-3795-8, p.118-119.
09 Topology derives from the Greek word for ‘place’.
11 Bit is the shortened form of binary digit, a unit of computer information equivalent to the result of the choice
between two alternatives as yes/no or on/off. Fuhrmann Peter, Bauplanung und Bauentwurf: Grundlagen und
Methoden der Gebäudelehre (1998), W. Kohlhammer GmbH Stuttgart, p.169.
Self-similarity is the symmetry that is scale-independent, where the smaller parts, elements are similar in the one
or other way to larger parts and to the whole.
14 Lynn Greg, Das erneuerte Neue der Symmetrie, Arch+ number 128 (1995), Arch+ Verlag GmbH Aachen,
ISSN 0587-3452, p.49.
ISSN 0587-3452, p.50.
ISSN 0587-3452, p.48.
18 The reason lies in the demand of city planning in form of an existing church to the right side, which Borromini
wanted to subordinate. Jencks Charles, Die Architektur des springenden Universums, Arch+ number 141 (1998),
Arch+ Verlag GmbH Aachen, ISSN 0587-3452, p.37.
19 Borcherdt Helmut, Architekten – Begegnungen 1956-1986 (1988), Georg Müller Verlag GmbH, ISBN
3-78844-2181-4, p.133.
20 Jencks Charles, The Architecture of the Jumping Universe, (1995), academy editions, ISBN 1 85490 406 X, p. 43.
21 Koch Wilfried, Baustilkunde – Das große Standardwerk zur europäischen Baukunst von der Antike bis zur
Gegenwart, Mosaik Verlag, p. 157, p. 168, p. 227.
22 Toman Rolf, Die Kunst der Gotik (1998),Könemann, ISBN 3-89508-313-5.
Gegenwart, Mosaik Verlag, p.148.
24 The word Gothic arises from the Italian ‘gotico’, which means barbarian, not-antique. The architectural expression
of the Gothic style is the pointed arch and the buttress with ribbed vault. Though both elements have been used
in the Romanesque style before – the pointed-arch has its origin in the Islamic world – only then they have been
favored before the many other known elements. Koch Wilfried, Baustilkunde – Das große Standardwerk zur
europäischen Baukunst von der Antike bis zur Gegenwart, Mosaik Verlag, p. 157, p. 168, p. 149.

26 Toman Rolf, Die Kunst der Gotik (1998),Könemann, ISBN 3-89508-313-5, p. 139.
28 Brockhaus, Der große Brockhaus Kompaktausgabe in 26 Bd., Band 8 Gasthörer bis Grimaud (1983) aktualisierte
18. Auflage, F.A. Brockhaus Wiesbaden, p.309.
29 ‘D’ is called the diameter-exponent. Mandelbrot Benoit B., Dr. Zähle Ulrich (editor of the German edition),
Die fraktale Geometrie der Natur (1991) einmalige Sonderausgabe, Birkhäuser Verlag Berlin, ISBN 3-
7643-2646-8, p.169.
30 For a long time Baroque meant pompous, which was also brought into literature and music – sometimes the word
‘overloaded’ can be added.
31 Brockhaus, Der große Brockhaus Kompaktausgabe in 26 Bd., Band 2 Archi bis Belgien (1983) aktualisierte 18.
Auflage, F.A. Brockhaus Wiesbaden, p.294.
33 Brockhaus, Der große Brockhaus Kompaktausgabe in 26 Bd., Band 18 Ratenwechsel bis Safanija (1983)
aktualisierte 18. Auflage, F.A. Brockhaus Wiesbaden, p.243.
The Portuese word ‘barocco’ means ‘little stone’, ‘irregular, sloping-round pearl’ which indicates the origin in the
artwork of jewellers.
34 Baedecker, Frankreich (1988) 4. Auflage, Karl Baedecker GmbH Deutschland.
35 Zerbst Rainer, Antoni Gaudí (1993), Benedikt Taschen Verlag GmbH, ISBN 3-8228-0067-8, p.190.
Gegenwart, quotation by Theo van Doesburg in the year 1924, Mosaik Verlag, p.386.
39 Jencks Charles, Die Architektur des springenden Universums, Arch+ number 141 (1998), Arch+ Verlag GmbH
Aachen, ISSN 0587-3452, p.66.
40 Brockhaus, Der große Brockhaus Kompaktausgabe in 26 Bd., Band 24 Weltkrieg bis ZZ (1983) aktualisierte 18.
Auflage, F.A. Brockhaus Wiesbaden, p.182.
41 Bovill Carl, Fractal Geometry in Architecture and Design (1996), Birkhäuser Bosten, ISBN 3-7643-
3795-8, p.127/128.
In general nature should not be imitated but the structure should be found and brought into architecture. This
structure is called the fractal concept, a basic idea on every scale, the variation of a theme, idea and rule, the
identification of parts and the whole, no strict symmetry.
43 Borcherdt Helmut, Architekten - Begegnungen 1956-1986 (1988), Georg Müller Verlag GmbH, ISBN 3-78844-2181-4, p.89.
Aachen, ISSN 0587-3452, p.106.
Aachen, ISSN 0587-3452, p.106.
ISSN 0587-3452, p.48.
Aachen, ISSN 0587-3452, p.59.
49 Edoardo Gellner, Alte Bauernhäuser in den Dolomiten, Verlag Georg D. W. Callwey (1989), ISBN 3-7667-0946-1.
50 Time is added because the family-background or money changes, which leads to additions or other space-
distributions. Then the existing house is the factor which determines that the changes remain under the same
concept. In Honkong there are high-rise houses where – illegal – additions are built like bird's-nests, which look
similar because of constructional limitations of depth, width and the use of timber. In this case the reason lies in
the very small space which the dwellings occupy.

54 The fluctuations of the fractal curve are statistically self-similar, which means that one part of the curve is not an
exact copy of other parts but they are statistically similar. To show this similarity one can zoom in on the curve, but
if the vertical and horizontal scaling factor is equal to one, then the resulting curve will be rougher than the original
curve. If a horizontal scaling factor of two is choosen instead, the curve will flatten out. Instead of that a properly
rescaling can be reached by changing the rescaling factor for the horizontal direction to the value of two and for
the vertical direction to the result of 2H. Bovill Carl, Fractal Geometry in Architecture and Design (1996),
Birkhäuser Bosten, ISBN 3-7643-3795-8, p.81/82.
55 This time it does not matter if the horizontal axis is streched or not because the vertical fluctuations
stay the same with the grid. Bovill Carl, Fractal Geometry in Architecture and Design (1996),
Birkhäuser Bosten, ISBN 3-7643-3795-8, p.82-87.
56 The algorithms of computer programs that generate fractal noise curves are approximations in their
simulations which lead to the fact that generated curves with a dimension above 1.5 will result in a
little bit lower measured dimensions, respectively curves generated to have fractal dimensions below
1.5 will offer higher measured dimensions. Bovill Carl, Fractal Geometry in Architecture and Design
(1996), Birkhäuser Bosten, ISBN 3-7643-3795-8, p.88.
On the one hand too much order does not bring up any surprise, which leads to the fact that the following parts
are known, that is determined, from the very beginning – music or architecture is then regarded as boring. On the
other hand too much surprise destroys any direction, its parts can be put together in any way, there is not any
sequence at all – man is upset and loses any kind of possible orientation.
58 Fuhrmann Peter, Bauplanung und Bauentwurf: Grundlagen und Methoden der Gebäudelehre (1998), W.
Kohlhammer GmbH Stuttgart, p.13.
6 Fine Arts and City Planning

01 Bruckgraber Iris and more, Kunst des 20. Jahrhunderts – Museum Ludwig Köln (1996) Originalausgabe, Taschen
Verlag GmbH Köln, ISBN 3-8228-8819-2, p.507.
04 von Lengerke Christa, Malerei heute – von Pollock bis Warhol, in: Walther Ingo F. (editor), Museum der Malerei,
Edition Atlantis, p.37.
05 Batty and Longley, Fractal Cities (1994), Academic Press Inc., ISBN 0-12-4555-70-5, p. 1.
The decisions on smaller scales may be the reason why ‘organically’ grown cities display irregularity in form
and do fit more comfortably in their environs.
07 Batty and Longley, Fractal Cities (1994), Academic Press Inc., ISBN 0-12-4555-70-5, preface p. VI.
On the other hand naturally grown cities often display regular parts because of man’s triumph over nature –
Baroque – or fast growth – turn of the 19th to the 20th century.
09 Müller Werner and Vogel Gunther, dtv-Atlas Baukunst, Bd.1 (12. edition 2000), Deutscher Taschenbuch Verlag GmbH
&Co.KG München, ISBN 3-423-03020-8, p. 109; ‘Hippodamian System’ p. 167; ‘Castrum romanum’ p.215.
11 Müller Werner and Vogel Gunther, dtv-Atlas Baukunst, Bd.2 (11. edition 2000), Deutscher Taschenbuch Verlag
GmbH &Co.KG München, ISBN 3-423-03021-6, p. 329.
12 ‘De Architectura’ 88-26 before Christ.
Koch Wilfried, Baustilkunde – Das große Standardwerk zur europäischen Baukunst von der Antike bis zur
13 Benevolo Leonardo, Die Geschichte der Stadt (8. edition 2000), Campus verlag GmbH Frankfurt/New York, ISBN
3-593-36439-5, p. 577.
16 Borcherdt Helmut, Architekten – Begegnungen 1956-1986 (1988), Georg Müller Verlag GmbH, ISBN 3-78844-2181-4, p.208.

17 Remember: Order alone is the lack of interesting turns, surprise, but in contrast only surprise cannot be understood,
which means that there is no confirmation, a city therefore needs both.
18 Borcherdt Helmut, Architekten – Begegnungen 1956-1986 (1988), Georg Müller Verlag GmbH, ISBN 3-78844-2181-4, p.214
20 The grey-brown concrete buildings have been ageing in a bad way, which is due to the different behavior of
material in a climate different from Europe.
Leaving the quarters of Le Corbusier means to return in the ‘normal’ living confusion of city-life with streets full of
people. Idealized city-plans based on grids have one important weak point, their lack of harmony with their
surroundings.
Borcherdt Helmut, Architekten – Begegnungen 1956-1986 (1988), Georg Müller Verlag GmbH, ISBN 3-78844-2181-4, p.210.
21 Borcherdt Helmut, Architekten – Begegnungen 1956-1986 (1988), Georg Müller Verlag GmbH, ISBN 3-78844-2181-4, p.101
25 Statement of the painter Ben Shahn. Arnheim Rudolf, Kunst und Sehen: e. Psychologie d. schöpfer. Auges
(expanded and revised edition 1978), Walter de Gruyter & Co., ISBN 3-11-006682-3, p. 93.
26 This approach reminds us of the statement of movement of Modern Architecture: ‘Form follows function’.
Batty and Longley, Fractal Cities (1994), Academic Press Inc., ISBN 0-12-4555-70-5, p. 42.
31 Batty and Longley, Fractal Cities (1994), Academic Press Inc., ISBN 0-12-4555-70-5, p. 228-229.
34 Various geophysical processes influence the forms of coastlines.
35 Batty and Longley, Fractal Cities (1994), Academic Press Inc., ISBN 0-12-4555-70-5, p. 180-181.
Maybe the kind of transportation system – tramway, underground, car, bus – and its availability – unbroken or
fragmentary – also changes fractal dimension.
7 Problems with Measuring

01 ‘Benoit version 1.2’ TruSoft Int’l, Inc. – Copyright (C) 1997, 1999; User manual.
02 http://www.mhri.edu.au/~pdb/fractals/fractdim (15.04.1999); User manual for a mac program.
8 Pictures
01 Jencks Charles, The Architecture of the Jumping Universe (1996), Academy Editions an imprint of Academy Group
Ltd, ISBN 1 85490 406 X, p.43.
06 http://stud4.tuwien.ac.at/~e9426503/soinfoges/mathematik.html (24.09.2001).
7003-9, p.43.
08 Fuhrmann Peter, Bauplanung und Bauentwurf: Grundlagen und Methoden der Gebäudelehre (1998), W.
Kohlhammer GmbH Stuttgart, p.102.

Foto-index
Batty and Longley, Fractal Cities (1994): picture 75, 76;
Bax Marty, Complete Mondrian (2001): picture 62, 63;
Benevolo Leonardo, Die Geschichte der Stadt (8. edition 2000): picture 06, 37, 65, 68, 71, 78;
Borcherdt Helmut, Architekten – Begegnungen 1956-1986 (1988): picture 68, 69, 71;
Bovill Carl, Fractal Geometry in Architecture and Design (1996): picture 28, 44, 48, 49, 55, 56, 58;
Chitham Robert, Die Säulenordnung der Antike (1994): picture 49;
Freytag & Berndt, Unterstufenatlas (1979): picture 01, 24, 25;
Gellner Edoardo, Alte Bauernhäuser in den Dolomiten (1989): picture 53, 54, 81, 83, 84; page 10; ‘9 Statistics’;
Gössel Peter and Leuthäuser Gabriel (editors), Frank Lloyd Wright (1994): picture 47;
Jencks Charles, Die Architektur des springenden Universums, Arch+ number 141 (1998): picture 04, 34, 50, 51, 52;
Jencks Charles, The Architecture of the Jumping Universe (1995): picture 04, 34, 50, 52;
Joedicke Jürgen, Architekturgeschichte d. 20. Jahrhunderts (1990): picture 55;
Koch Wilfried, Baustilkunde: picture 40, 41, 65, 66, 67;
Koepf Hans, Baukunst in 5 Jahrtausenden (10. edition 1990): picture 05, 41, 49;
Lynn Greg, Das erneuerte Neue der Symmetrie, Arch+ number 128 (1995): picture 31, 32;
Müller Werner and Vogel Gunther, dtv-Atlas Baukunst, Bd.1 (12. edition 2000): picture 65;
Pevsner, Honour, Fleming, Lexikon der Architektur (1992): picture 46;
Pothorn Herbert, Das große Buch der Baustile (1998): picture 05;
Seibt Ferdinand, Glanz und Elend des Mittelalters (1987): picture 03;
Storrer William Allin, The Frank Lloyd Wright Companion (1993): picture 58;
Toman Rolf, Die Kunst der Gotik (1998): picture 37, 38, 39, 40;
Zerbst Rainer, Antoni Gaudí (1993): picture 42;
pictures by my own: CAD grafics: picture 03, 15, 16, 19, 20, 24, 25, 29, 30, 33, 36, 40, 44, 45, 54, 57, 58, 59, 60, 61,
66, 67, 77, 79, 80, 82; boxcountings with CAD: picture 43, 44, 48, 49, 53, 56, 83;
other grafics: picture 02, 08, 09, 10, 11, 12, 13, 18, 21, 22, 23, 28, 72, 74;
photographs: picture 26, 70;
programs: Benoit.exe; Benoit 1.2, TruSoft International Inc. (1996): picture 61; ‘9 Statistics’;
Chaos.exe; Visions of Chaos – version 33.8 by J. Rampe (2001): picture 18, 21;
ChaosPro.exe; Random Growth – DLA Ver. 1.1 (2000): picture; 13
FractDim.exe; Program to calculate fractal dimension by M.Laubscher: ‘9 Statistics’;
Fe.exe; Fractal Explorer 1.22se (2001): picture 17;
Fractal Dimension Calculator by Paul Bourke (1993): ‘9 Statistics’;
FrMesh.exe; by Paul McGuire (1993): picture 20;
LMay2.exe; Lindenmayer-Fraktale 2.1 by Ulrich Schwebinghaus (1999-2001): picture 19;
Spsswin.exe; spss for Windows (1996): ‘9 Statistics’;
Winfract.exe; Winfract version 18.21 – The Stone Soup Group (1990,1993): picture 07, 14;
VMbrot.exe; by David A. Cromley (1994): picture 14;
Websites: www3.justnet.ne.jp/~wadayoshio/indexe.htm (29.11.2001): picture 27;

www.archinform.de/projekte/600.htm (17.06.2002): picture 04;
www.globalgallery.com/images/bm-p635.jpg (17.06.2002): picture 64;
www.greatbuildings.com/buildings/Ozenfant_Studio.html (26.07.2002): picture 45;
www.islamabad.net/guidemaps.htm (26.02.2002): picture 69;
www.joho-kyoto.or.jp/KICH/index_e.html (13.06.2002): picture 35;
www.jpl.nasa.gov/radar/sircxsar/gwall.html (29.11.2001): picture 27;
www.lib.utexas.edu/maps/historical/vienna_1858.jpg (26.11.2002): picture 72;
www.mscs.mu.edu/~mikes/ArtsWeb/goff/bavphot.gif (17.06.2002): picture 04;
www.papyrusversand.de/Herkunft/herkunft.html (29.11.2001): picture 46;
www.wien.gv.at/stadtentwicklung/05/03/04/11.htm (26.07.2002): picture 77;
www.wien.gv.at/wiengrafik/suche.htm (23.08.2002): picture 73;

Fractals and Fractal Architecture

Uploaded by

Document Informationclick to expand document informationFRACTALS

Document Informationclick to expand document information

Copyright:

Available Formats

Fractals and Fractal Architecture

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fractals and Fractal Architecture

Uploaded by

Copyright:

Available Formats

See

Fractals and Fractal Architecture

THESIS · JANUARY 2003

Available from: Wolfgang E. Lorenz

“Fractals and Fractal Architecture”

ausgeführt zum Zwecke der Erlangung des

O.Univ.Prof. DI Dr.techn. Georg FRANCK-OBERASPACH

eingereicht an der technischen Universität Wien

Fakultät für Architektur und Raumplanung

2 Fractals – A Definition page 8

3 Different Fractals page 16

Vienna, 2002 – page 2– Wolfgang E. Lorenz

5 Fractals and Architecture page 29

Vienna, 2002 – page 3– Wolfgang E. Lorenz

6 Fine Arts and City Planning page 53

7 Problems with Measuring page 63

9 Statistics page 121

Vienna, 2002 – page 4– Wolfgang E. Lorenz

As the following pages indicate, fractal geometry, in opposition to Euclidean geometry,

1.1 Mandelbrot – Fractals – Theories and Explanations

Vienna, 2002 – page 5– Wolfgang E. Lorenz

Euclidean – fractal geometry

Vienna, 2002 – page 6– Wolfgang E. Lorenz

Vienna, 2002 – page 7– Wolfgang E. Lorenz

2.1 What is a Fractal?

2.2.1 Characteristics – A Fractal is Rugged

Vienna, 2002 – page 8– Wolfgang E. Lorenz

Another example of scale-dependence is the length-measurements of borders between

Vienna, 2002 – page 9– Wolfgang E. Lorenz

unit length 100

As unit-lengths lead to different total-lengths, borders, coastlines and other

2.2.2 Characteristics – A Fractal is Self-Similar

2.2.2.a Self-Similar and Self-Affine

Any structure is self-similar if it has under-gone a transformation in which the proportions

Vienna, 2002 – page 10– Wolfgang E. Lorenz

On looking at the border of a Euclidean object such as a circle, something different

2.2.2.c Self-Similar Structures in Architecture

Vienna, 2002 – page 11– Wolfgang E. Lorenz

2.2.2.d Self-Similar Structures and Cities

2.2.2.e Other Disciplines

Vienna, 2002 – page 12– Wolfgang E. Lorenz

2.2.3 Characteristics – A Fractal is Infinitely Complex

2.2.4 Characteristics – A Fractal is Developed through Iterations

Self-similarity, as described before in this chapter, can be produced by iterations, which

A film-camera and a television viewer in a dark room can illustrate feedback,

Vienna, 2002 – page 13– Wolfgang E. Lorenz

2.2.5 Characteristics – A Fractal Depends on Starting Conditions

2.2.6 Characteristics – A Fractal is Common in Nature

Vienna, 2002 – page 14– Wolfgang E. Lorenz

differences natural – ture fractals

Vienna, 2002 – page 15– Wolfgang E. Lorenz

3.1 The ‘true’ Mathematical Fractals

3.1.1 Cantor Set[01]

Vienna, 2002 – page 16– Wolfgang E. Lorenz

This sounds quite similar to the distribution of centers or the distribution of

3.1.2 Sierpinski Gasket[04]

3.1.3 Koch Curve[05]

Vienna, 2002 – page 17– Wolfgang E. Lorenz

r1 = r0/2; L1 = N1r1; r2 = r1/2; L2 = N2r2;